DOS Coders1 Message Board ,----------------------------------------------------------------------------- | | Arachne DOS Coders1 Message Board | | | | | | | | [ For the Operational discussions of the Arachne C Coders ] | | | | ------------------------------------------------------------+-------+---+ | | ============================================================================ | | Sat, Apr 02, 2005 - 6:30:06 am | | mht> | We had some W.B. cartoons. I don't remember much of them -- maybe | there was Marvin as well and I just forgot... | | So, Ray, when do you celebrate your birthday? | | | ============================================================================ | | Sat, Apr 02, 2005 - 8:57:45 pm | | ray> | Michal, | | It's a long way off, I was just BSing ;-) | | Besides, what I reeeealy want for my birthday is some | fonts that use IBM characters :-) | | | | ============================================================================ | | Mon, Apr 11, 2005 - 8:16:03 am | | glennmcc> | They got it back up. :)) | | ============================================================================ | | Mon, Apr 11, 2005 - 12:13:19 pm | | gregy> | Yep...worked hard over the weedend.... | | Now, I'll try to get that new BoardCORE (that dislays spaces and TABs) | installed and running... | | | | ============================================================================ | Š| Mon, Apr 11, 2005 - 9:51:15 pm | | ray> | She's back. (Good news about those spaces!) ... Hmmm, can you | process tabs? IINM, those will always pop you out of the window, no? | | | | ============================================================================ | | Mon, Apr 11, 2005 - 10:50:42 pm | | gregy> | hmmm..."pop you out of the window, no?".....I do NOT understand...do you | mean that using Glenn's "new model" board source, the posts will lap | over outside the right margin....??? | | | ============================================================================ | | Tue, Apr 12, 2005 - 1:12:07 pm | | ray> | Na, it's just that the tab always moves you to another field, right? | I'm typing now, but if I hit the tab I go the the "Nick" line then have | to tab twice to get back into the .... but of course that's irrelevant | to how a imported block of text might be rendered isn't it? | | | ============================================================================ | | Tue, Apr 12, 2005 - 8:01:21 pm | | gregy> | ahhh...I see what you mean....I misspoke....I meant "tabs" in the | posts....which are NOT actual tabs...they will be spaces (I think, I | don't have one to play with, until I get Glennmcc's new code installed | here)....but...anyway....it will enable you guys to post code with the | indentation preserved.. | | | | ============================================================================ | | Wed, Apr 13, 2005 - 10:42:11 pm | | ray> | Good! The board is the place for small snips. | | | ============================================================================ | | Thu, Apr 14, 2005 - 2:28:12 am | | gregy> Š| Well, yes....and also the place for LARGE problems...which it now seems | we may have. | | After thinking about Geoff's ISP's new DNSserver numbers problems...I | went and subbed his new DNS numbers into ARACHNE.CFG and tried again to | connect....his new DNSservers refused to answer my Arachne 1.71... | .. | Then I fired up my Win3.1, plugged his new DNS numbers into the standard | Win3.1 DialUpNetwork Client, dialed and connected, then launched my | Netscape Communicator 4.08b, clicked my Bookmarks' Google link,.... | ...and....it got a good return from those new DNSes and connected right | up....finest kind..... | ... | So, we may have a problem..... | Would any of you guys running any OperatingSys other than DOS, try | those numbers....??? | They are: | ;194.176.65.5 | ;195.97.231.3 | | | ============================================================================ | | Thu, Apr 14, 2005 - 2:29:46 am | | gregy> | Please report back what OpSys and Browser you used.... | .. | Thanks... | | | | ============================================================================ | | Thu, Apr 14, 2005 - 2:52:18 am | | gregy> | In fact, had a SECOND Wild Thought...!!! | ... | Will any of you guys please try these new DNS numbers in some DOS | Browsers OTHER THAN Arachne.....??? DOSLynx, Lynx, Links, Bobcat, | etc....??? | | Please report back with the DOS you used and what Browser... | | Thanks... | | | ============================================================================ | | Thu, Apr 14, 2005 - 3:15:42 am | | gregy> | Furthermore....I just went back and tried all these: | Win3.1 Š| ------ | Netscape Communicator v. 4.08b | Netscape Navigator v. 3.04 | Internet Explorer v. 3.03d | mIRC v. 5.71 | | ...and they all worked with Geoff's ISP's new DNSserver numbers... | ... | :(((( | | | ============================================================================ | | Thu, Apr 14, 2005 - 7:38:49 am | | ray> | Hey, this is good news, not bad. It shows we have a simple bug, not | some new fangled 'standard' to deal with. | | | | ============================================================================ | | Thu, Apr 14, 2005 - 6:50:10 pm | | glennmcc> | Gregy, | | Are you sure that you have 'dynamic DNS' disabled in Win3.1 | | As you'll see in my reply to the mailing list.... I disabled dynamic DNS | in the Win98 (1st edition) DUN and plugged-in Geoff's DNSs | | Netscape... no go | MSIE....... no go | FireFox ... no go | | I then did the same in Slackware Linux v7 | | lynx....... no go | links...... no go | ftp........ no go | telnet..... no go | Netscape... no go | | | ============================================================================ | | Thu, Apr 14, 2005 - 10:48:50 pm | | ray> | 0 | 1 | 2 | 3 Š| (guess what I'm looking for ;-) | | | ============================================================================ | | Thu, Apr 14, 2005 - 10:58:56 pm | | gregy> | No, Glenn...I'm sure that I do NOT have "Dynamic DNS" disabled...unless | that is the default. I didn't see such an option on that Win3.1 Shiva | DUN's Options--->Properties--->TCP/IP page. I'll go look around for | it. | | That may be why my Win3.1 DUN connected, even with those djbdns DNS | numbers plugged into the DUN Options page.... | | I'll get back to you with the results... | | | | ============================================================================ | | Fri, Apr 15, 2005 - 2:33:49 am | | gregy> | Glenn....nope....that Win3.1 DUN has no "Dynamic DNS" anywhere around | in its setup pages. | I noticed the "About" page cites ShivaCorp and FTPCorp as the authors. | FTP being responsible for the WinsockDLL and TCP...so...I wonder it | FTP somehow lucked out and "just happened to" write their TCP stack | to be in compliance with this djbdms "new" protocol.....hmmmm...not | likely, but, then, it shouldn't be likely that it would work with | those DNSes when nothing you have will work....even your Linux apps. | ... | Strange, and needs more investigation....sumpin ain't making sense, | here... | | | ============================================================================ | | Fri, Apr 15, 2005 - 12:26:20 pm | | glennmcc> | As far as I know.......... DJBDNS did not exist that long ago. | | More likely, what's happening is that it _is_ using 'dynamic DNS' | in additon to the 'plugged-in' DNSs. | | Something like so.... | | At login time, it gets the DNSs assigned bt the ISP being logged into. | | It will first attemt top get the needed IP address via the 'plugged-in' DNSs | | If that fails, it then tries the DNSs which were obtained at login time Š| via dynamic assignment. | | Just as-if we had 4 DNSs in WATTCP.CFG | | NameServer Geoff's 1st DNS | NameServer Geoff's 2nd DNS | NameServer Gregy's 1st dynamic DNS | NameServer Gregy's 2nd dynamic DNS | | If this be the case, this would be what's happening. | | NameServer Geoff's 1st DNS (fails to reply with requested IP) | NameServer Geoff's 2nd DNS (fails to reply with requested IP) | NameServer Gregy's 1st dynamic DNS (replies with requested IP) | NameServer Gregy's 2nd dynamic DNS (does not get contacted since we already | a reply) | | | ============================================================================ | | Sun, Apr 17, 2005 - 5:48:23 pm | | ray> | 0 | 1 | 2 | | | | ============================================================================ | | Sun, Apr 17, 2005 - 9:12:32 pm | | gregy> | Glenn...yep...that's something to think about.....BUT...ref my list | post, my ISP's support people say NO such thing as "dynamic DNS | assignment - all static" as far as their servers are concerned..AND | they have never heard of "dynamic DNS" (doesn't mean it doesn't exist, | just they're not using it). | | Ref my list post, if you see anything I'm doing wrong, or missing, let | me know. We can try that out. I actually went back into that, today, | figuring that I WOULD find something I was doing/not doing that would | give me the lookup/response on good servers, instead of the servers I | had poked into the ShivaDialer. Thought I'd found it with that error | on the ".3" instead of ".31"...but...nope....it still works...and, as | you can tell from my post, I thought of several other things that might | be causing my Win3.1 to still connect, even with Geoff's new DNSserver | numbers poked into the Dialer.... | Still mystifies me....but...if my Win3.1 DUN is actually connecting, | then it might be a way of checking to see what kind of packets those | DJBDNS servers are accepting and responding with. | It's not an immediate threat...just something we need to check into | some more....so we're prepared if we need to be. | Š| ============================================================================ | | Wed, Apr 20, 2005 - 12:46:44 pm | URL | | glennmcc> | Experimental core uploaded. | | http://www.cisnet.com/glennmcc/my-stuff/core1234.zip | | | ============================================================================ | | Wed, Apr 20, 2005 - 12:53:20 pm | | glennmcc> | Btw, | | In addition to using my fresh compile of WATTCPLG.LIB | set to port 1234 instead of 997 | | It also has......... | | --- clipped from my current changes.txt --- | | 1.85;GPL,386+ stable for DOS | ---- | Release date: ??? ??, 2005 | | Mar 06, 2005 | Fixed a problem with the & symbol contained in URLs | http://www.cisnet.com/glennmcc/testing_&_symbol/&-amp-bug.htm | All 7 links now work perfectly. | | Mar 06-14, 2005 | Ray and I fixed a few more glitches with CSS link hover | when used within a framed page. | | Mar 23, 2005 | Fixed yet another glitch with CSS link hover. | | Apr 08, 2005 | Fixed the problem with 'broken FTP dir listings' at such places | as ftp://ftp.cdrom.com/.2/simtelnet/ | | | | ============================================================================ | | Wed, Apr 20, 2005 - 5:04:52 pm | | ray> | It's a pleasure having to bitch about so much progress to keep up | witth :-) | Š| | ============================================================================ | | Fri, Apr 22, 2005 - 1:44:37 pm | | ray> | Joe, | | Built that new .lib in, and for the record, it only cost me an | extra 64 bytes :-) | | | ============================================================================ | | Fri, Apr 22, 2005 - 4:05:07 pm | | glennmcc> | I was already using the copy of the lib from Nov 2004 so | | | ============================================================================ | | Fri, Apr 22, 2005 - 4:10:17 pm | | glennmcc> | Oops, hit the F2 key by mistake | | I was already using the copy of the lib from Nov 2004 so | mine remained exactly the same size. | | Joe's new LIB is exactly the same size as that one. | | 163,840 03-16-01 09:23a | e:\arachne\asrc\asrc185\changes\wattcp.lib\2001\wattcplg.lib | 159,232 11-02-04 10:25p | e:\arachne\asrc\asrc185\changes\wattcp.lib\2004\wattcplg.lib | 159,232 04-22-05 01:06a | e:\arachne\asrc\asrc185\changes\wattcp.lib\joe\wattcplg.lib | total files 3 total bytes 482,304 disk free space 54,026,240 | | | ============================================================================ | | Fri, Apr 22, 2005 - 4:19:34 pm | | glennmcc> | Here's something strange. | | I just now built a fresh one using Joe's modified files. | | 152,064 04-22-05 05:11p e:\wattcp\wat1104\lib\wattcplg.lib | 159,232 11-02-04 10:25p e:\wattcp\wat1104\lib\original\wattcplg.lib | total files 2 total bytes 311,296 disk free space 52,756,480 | | Why is the fresh build 7,168 bytes smaller ? Š| | I built it by simply CDing to the SRC dir and typing 'make' | | 'make' of-course built it using Borland C/C++ v3.1 | | | | ============================================================================ | | Fri, Apr 22, 2005 - 4:27:49 pm | | glennmcc> | Hmmmmmm, Something is wrong. | | I just tried to recompile core.exe using my build of the LIB | | Linking CORE.EXE: | Linker Error: Undefined symbol _set_timeout in module PCTCP | Linker Error: Undefined symbol _chk_timeout in module PCTCP | Linker Error: Undefined symbol _intel in module PCTCP | Linker Error: Undefined symbol _set_ttimeout in module PCTCP | Linker Error: Undefined symbol _inchksum in module PCTCP | Linker Error: Undefined symbol _intel16 in module PCTCP | Linker Error: Undefined symbol _qmove in module BSDNAME | | | | | ============================================================================ | | Fri, Apr 22, 2005 - 5:01:23 pm | | glennmcc> | curiouser and curiouser. | | I just did a new build of the LIB after placing a copy of the original | back into the \LIB\ subdir. | | The new build is now larger than my 1st build when the \LIB\ subdir was empt | | 159,744 04-22-05 05:38p e:\wattcp\wat1104\lib\wattcplg.lib | 159,232 11-02-04 10:25p e:\wattcp\wat1104\lib\original\wattcplg.lib | total files 2 total bytes 318,976 disk free space 52,707,328 | | Core.exe now compiles without errors but is still exaclty the same size | as with Joe's copy of the LIB | | 441,464 04-22-05 4:07p l:\arachne\core.exe | 441,464 04-22-05 5:39p c:\arachne\core.exe | | | And the curiousities continue...... | | Why are these 2 bytes different in the 2 files ??? | Š| Comparing files in... | C:\ARACHNE | with files in... | L:\ARACHNE | | Comparing CORE.EXE with CORE.EXE - Compare failure. | Offset 3309eh Source = 40h Destination = dch | Offset 3309fh Source = 2h Destination = 5h | 2 Mismatches - No more mismatches were present. | | | ============================================================================ | | Fri, Apr 22, 2005 - 5:10:48 pm | | glennmcc> | One last test, this time starting-off with Joe's LIB in the \LIB\ subdir. | | Exactlty the same results as the ladt test with the Nov 2004 LIB in there. | | | ============================================================================ | | Fri, Apr 22, 2005 - 6:15:35 pm | | ray> | I'll wait for Joe's comments on this, for now I've just taken his | .lib as he supplied it and all seems OK. | | | | ============================================================================ | | Fri, Apr 22, 2005 - 6:30:58 pm | | glennmcc> | OK, I found the reason for the difference in filesize of the LIB itself. | | I renamed the dir where Borland is installed. | | rendir c:\borlandc c:\bc31 | | Changed my path statement in autoexec,bat to point there and rebooted. | | My new WATTCPLG.LIB is now the same filesize as Joe's. | | 159,232 4-22-05 7:09p e:\arachne\asrc\asrc185\wattcplg.lib | 159,232 4-22-05 1:06a e:\arachne\asrc\asrc185\changes\wattcp.lib\joe\wattcpl | | | But the compare of the 2 copies of core.exe still show those same 2 bytes | different. | | ??????? | Š| And the LIB files temselves have 18 different bytes. | | L:\>comp /m:0 e:\arachne\asrc\asrc185\wattcplg.lib | e:\arachne\asrc\asrc185\changes\wattcp.lib\joe\wattcplg.lib | | Comparing files in... | E:\ARACHNE\ASRC\ASRC185 | with files in... | E:\ARACHNE\ASRC\ASRC185\CHANGES\WATTCP.LIB\JOE | | Comparing WATTCPLG.LIB with WATTCPLG.LIB - Compare failure. | Offset 3beh Source = 31h Destination = 32h | Offset 3bfh Source = 9ah Destination = 99h | Offset 68dh Source = 31h Destination = 32h | Offset 68eh Source = 9ah Destination = 99h | Offset 7cah Source = 31h Destination = 32h | Offset 7cbh Source = 9ah Destination = 99h | Offset 90bh Source = 31h Destination = 32h | Offset 90ch Source = 9ah Destination = 99h | Offset a3ch Source = 31h Destination = 32h | Offset a3dh Source = 9ah Destination = 99h | Offset baah Source = 31h Destination = 32h | Offset babh Source = 9ah Destination = 99h | Offset d0bh Source = 31h Destination = 32h | Offset d0ch Source = 9ah Destination = 99h | Offset e6bh Source = 31h Destination = 32h | Offset e6ch Source = 9ah Destination = 99h | Offset fbeh Source = 31h Destination = 32h | Offset fbfh Source = 9ah Destination = 99h | 18 Mismatches - No more mismatches were present. | | Hmmmmm, 9 occurences of the same pair of differences. | | mine 31h Joe's 32h ( 1 smaller in mine ) | | mine 9ah Joe's 99h ( 1 larger in mine ) | | Decimal | | 49 50 | | 154 153 | | ASCII | | 1 2 | | š ™ | | | | ============================================================================ | | Fri, Apr 22, 2005 - 7:58:44 pm | Š| glennmcc> | Ah ha....... | | mine........ Turbo Assembler Version 3.1š | | Joe's ...... Turbo Assembler Version 3.2™ | | | ============================================================================ | | Fri, Apr 22, 2005 - 8:13:51 pm | | glennmcc> | Google to the rescue once again. :) | | Found and downloaded tasm v3.2 | | Now we'll see what happens. | | | ============================================================================ | | Fri, Apr 22, 2005 - 8:14:21 pm | URL | | glennmcc> | Oh yeah, here's the place I found it. | | http://www.singlix.org/trdos/tasm.zip | | | ============================================================================ | | Fri, Apr 22, 2005 - 9:22:28 pm | | glennmcc> | This is screwy as hell. | | Now I get 21 differences strewn all throgh the new LIB file and none of | them have to do with the version number. | | Screw it........ | I'm gonna use Joe's build of wattcplg.lib and leave it at that. :) | | ============================================================================ | | Fri, Apr 22, 2005 - 10:37:48 pm | | ray> | Yup, sometimes you are better off just trusting that your coworkers | know what they are doing in their own departments. BTW have you guys | checked out Michal's alternate system font? He was going to upload it | to Yahoo. IMHO, it's a clear improvement -- of course these things | are subjective. | Š| | | ============================================================================ | | Sat, Apr 23, 2005 - 7:26:04 am | | Joe> | Hi folks (haven't been here for a while, been busy battling with | some other horrors of a Sun/Solaris bent;-), | | I'll have to check my build of wattcplg.lib some time. I know I also | kept getting strange behaviour by the Borland tools, I just persisted | until it all worked! I'll let you know if I find anything ... | I don't think there'll be any problem ... I've been using an updated | core for a few days now and all seems fine (bearing in mind I didn't | have any DNS problems to begin with). | | With any luck, Erick may provide some feedback too (you may have | noticed that I CC'ed him on that AD posting). I'll let you know if | I hear anything. | | | ============================================================================ | | Sat, Apr 23, 2005 - 8:47:36 am | | ray> | Ya, you did the right thing CCing Erick, I don't know how much interest | he still has in his baby, but that's for him to decide. If there are | bugs in your .lib all we've gota do it revert back to the old .lib | in case of some trouble or other until the bug is shaken out. | | | Too bad about the Solaris :-( | | | ============================================================================ | | Mon, Apr 25, 2005 - 5:40:45 pm | | ray> | Day two of Arachne without MHT :-( | | | | ============================================================================ | | Mon, Apr 25, 2005 - 5:47:47 pm | | glennmcc> | It ain't gonna be as much fun anymore now that we must go on without him. :- | | ============================================================================ | | Tue, Apr 26, 2005 - 1:09:06 am Š| | gregy> | Where'd Michal go.....??? | | | ============================================================================ | | Tue, Apr 26, 2005 - 7:26:11 am | | ray> | He's quit :-( | Not enough time, too many other projects ... | | | | ============================================================================ | | Tue, Apr 26, 2005 - 10:09:43 pm | | ray> | Hey Glenn, | | I'm just noticing that some config keywords like "PhoneNumber" are | irrelevant when using lsppp arguments to make a connection. Commenting | these out thus does no harm, OTOH, is is spooky to see all the | fields in the dial page blank. Leaving the kewords in place just | for visual comfort ain't my style, since those keywords wouldn't be | doing anything. Thoughts? BTW I noticed this while practicing | starting Arachne with the contents of CONFIG.LOG used in place of | ARACHNE.CFG (as a diagnostic tool), which works quite nicely. | The idea is that if you use the CONFIG.LOG produced by ARACHNE.CFG | *as* ARACHNE.CFG, the next CONFIG.LOG produced should be identical | to the previous one if all your config options are valid. | | ============================================================================ | | Wed, Apr 27, 2005 - 5:19:17 am | | Joe> | Hmmm ... It depends how you use LSPPP ... for me (using LSPPPCFG | of course), PhoneNumber *is* used. | | | ============================================================================ | | Wed, Apr 27, 2005 - 5:35:13 am | | Joe> | Oh, BTW, it looks like you were right about Erick having lost | interest in WatTCP - no response from him so far. | | Also, I've reviewed the build of wattcplg.lib, everything still | looks fine. I presume users with "finicky" DNSes are now OK too? | | That Sun/Sparc/Solaris monster is still not co-operating. I've Š| got a bunch of hardware to play with, but I'm sure some of it's | faulty. At the "BIOS" level, called "OpenBoot", the thing is a | Forth interpretor - unfortunately I don't know Forth, so it's | quite confusing and complicated, not at all like a PC. :-/ | | | | ============================================================================ | | Wed, Apr 27, 2005 - 5:49:50 pm | | ray> | Joe, | | RE: lspppcfg: That served me very well until I decided to try the | command line approach. I tend to like to do things in the most | direct manner, but the above could be reason enough to keep using | lspppcfg. One question: if you manualy edit, say, your phone number | in the dial page, will the new number be used if you hang up and | reconnect, or do you have to restart for that to take effect? In | other words, is lspppcfg smart enough to keep tabs on the variable, or | is it just read once at startup? If the former, then it's back to | lspppcfg for me. | | | | ============================================================================ | | Wed, Apr 27, 2005 - 8:07:44 pm | | glennmcc> | Having lsppp.cfg in the current dir 'over-rides' the command line options. | | The options are taken from the cfg instead. | | In addition..... | lsppp.exe can't take its options from the other lins in arachne.cfg | | Only miniterm.exe can do that. | | Therfore in order to use lsppp.exe, we must either specify all of the | command line options on the lsppp.exe line in arachne.cfg | | Dialer @lh lsppp.exe /x /i:4 /b:bx800 /d:330-423-4005 /B:115200 /V:60 | /U:glennmcc@cisnet.com /P:*****>PPP.LOG | | Or use just this line and then have all of the options in lsppp.cfg | | Dialer @lh lsppp.exe>ppp.log | | If we use the 1st line _and_ lsppp.cfg is in the arachne_main dir | then the end result is the same as the 2nd line. | | The only way to use the 1st line is _not_ have lsppp.cfg in the arachne_main | Š| ============================================================================ | | Wed, Apr 27, 2005 - 9:38:43 pm | | ray> | Glenn, | | Rodger that with lsppp itself but Joe's lspppcfg.exe is, I suspect, | capable of, or could be made capable of, reading arachne.cfg lines. | It makes things less than clear when 'lsppp.cfg' and lspppcfg refer | to different things of course. | | | ============================================================================ | | Wed, Apr 27, 2005 - 10:06:21 pm | | glennmcc> | Perhas, but as it is right now lspppcfg.exe is only setup to use | lsppp.exe as the packet driver and not as the dialer. | | It leaves miniterm.exe in-place as the dialer and simply replaces | epppd.exe with lsppp.exe as the packet driver. | | By using the dialer line I showed above and this connection line.... | | Connection @lsppp.exe /?>NUL | | lsppp.exe is now used as _both_ the dialer _and_ the packet driver. | | Now both epppd.exe and miniterm.exe can be dumped because niether of | them is needed anymore. | | | ============================================================================ | | Wed, Apr 27, 2005 - 10:12:39 pm | | glennmcc> | In addition, none of these lines in arachne.cfg are needed anymore. | | Port | Irq | Base | Mode | Speed | InitString | DialString | PhoneNumber | WaitFor1 | Response1 | WaitFor2 | Response2 | WaitFor3 | Response3 Š| PPPusername | PPPpassword | | [tcp/ip] | NameServer | AltNameServer | | All of those can be deleted and everthing still works fine because | they are only used by either epppd.exe or miniterm.exe and not by | core.exe nor by lsppp.exe | | ============================================================================ | | Wed, Apr 27, 2005 - 10:17:13 pm | | Joe> | Hi Ray, | | Well ... it's exactly per Arachne/Miniterm/EPPPD ... remember, with | LSPPPCFG, the dialling is done by Arachne/Miniterm, only the PPP | is done by LSPPP. At any time, changing the PhoneNumber or account | details via pppenhan.htm (IIRC) or arachne:pppsetup or the PPP wizard | or editing arachne.cfg or loading a new ACF (have I remembered them | all?;-) will make the Arachne/Miniterm/LSPPPCFG/LSPPP combination | use the current (new) setting whenever you redial (IIRC). | | | ============================================================================ | | Wed, Apr 27, 2005 - 10:54:20 pm | | glennmcc> | Hi Joe, | | Do you think it might be worth-while and worth your time to modify your | lspppcfg 'package' so that lsppp will also be used as the dialer ? | | ie: cfg-ppp lh lsppp would place these 2 lines into arachne.cfg | | Connection @lsppp.exe /?>NUL | Dialer @lh lsppp.exe>ppp.log | | It would also grab the needed lines from arachne.cfg and place them into | lsppp.cfg | | In my case, this is what it would do. | | --- from arachne.cfg --- | Irq 4 | Base 0xb800 | Speed 115200 | InitString ATf50 | PhoneNumber 330-423-4005 | PPPusername glennmcc@cisnet.com | PPPpassword quark Š| _________________________ | | --- into lsppp.cfg --- | /i:4 | /b:bx800 | /B:115200 | /M:ATf50 | /d:330-423-4005 | /U:glennmcc@cisnet.com | /P:***** | ______________________ | | As we see there is a stumbling block. | | For my 'non-standard' port address.... | Arachne/miniterm needs 0xb800 where-as lsppp.exe needs bx800 | | ============================================================================ | | Thu, Apr 28, 2005 - 4:21:21 pm | | ray> | Yeah, if the old keywords are obsolete, let's forget them or make them | work again. I'll bet Joe could reconfigure lspppcfg to make it | dial as well as everything else which, as Glenn says, gets rid of | miniterm and epppd. So then either lspppcfg would process the existing | Arachne keywords or those keywords would be dropped and the wizards | and various input boxes that use them would need to be rewired to | get their values from lsppp.cfg. The former has better continuity if | Joe wants to do it. | | | ============================================================================ | | Thu, Apr 28, 2005 - 4:24:09 pm | | ray> | ... or, we could add a bit of code into core which makes its own | lsppp.cfg from the existing keywords. Not a big job. | | | | ============================================================================ | | Thu, Apr 28, 2005 - 6:22:02 pm | | ray> | Hmmm, there's a whole lot of stuff related to this issue. | 'Course I don't understand it fully yet, but looking in ArachneDial() | it seems that 'pppdrc.cfg' isn't needed if lsppp.exe is used, and | 'UseTerminal' and 'TerminalWindow' are only of use if miniterm is | used/usable which in turn seems to mean that the old redundant keywords | Glenn mentioned above could be needed just for 'terminal' unless | that too is scrapped which would result in further simplifications | of the keywords and the code in ArachneDial(). Š| | ============================================================================ | | Fri, Apr 29, 2005 - 3:10:08 am | | Joe> | Hi all, | | Certainly, it would be relatively easy for LSPPPCFG to read the | appropriate stuff from arachne.cfg and use LSPPP for dialling. | However, I chose not to do this because 1) I wanted to retain the | terminal window capability (hi Ray;-) and 2) Because the LSPPP | dialler is less robust than Miniterm's (I have had complaints | about this from some users that tried to use LSPPP for everything). | | And, if I was convinced that nobody wanted the terminal window stuff | and that the LSPPP dialling was 100%, and changed LSPPPCFG to suit, | then most of the keywords above would still be required anyway. BTW, | presently, LSPPPCFG does use 'pppdrc.cfg', and whatever other similar | files are defined in the Arachne doc's (can't recall the details just | now, but as I recall there's an alternative file name and path search | options involved). Basically, LSPPPCFG uses the same files and settings | as EPPPD does, if that puts things in perspective. | | | ============================================================================ | | Fri, Apr 29, 2005 - 3:18:30 am | | Joe> | Hi Glenn, | | I forgot to mention ... I think you'll find that lsppp.exe doesn't | _require_ "bx800" syntax for its port address, merely it has poor | syntax checking and accepts this as "0xb800". You should find that | the correct syntax (ie. 0xb800) works just fine in both cases. | | | | ============================================================================ | | Fri, Apr 29, 2005 - 3:24:32 am | | Joe> | Oh, about the WaitFor1, Response1, etc. keywords ... AFAIK, these | are used only for SLIP connections, right? (I certainly don't use | them with either EPPPD or LSPPP.) | | ============================================================================ | | Fri, Apr 29, 2005 - 4:01:10 am | | gregy> | Question I have....since sometimes we (Wizzy, and maybe others) want | to dialup/connect/install packet driver...without loading Arachne (i.e. Š| I mean installing Arachne as "Online" instead of "dialup"....and we | know miniterm/epppd will do that, bacause Wiz has a package that does | that....since all of these things..... | ...can LSPPP be configured to do this, also.........???? | | | ============================================================================ | | Fri, Apr 29, 2005 - 8:02:10 am | | phantom> | Hi Gregy, | | Not sure what this "without loading Arachne" stuff is about, | particularly how dial-up is initiated, but the equivalent | would probably be something like : | | miniterm | lspppcfg | lsppp | | You'd probably do this via a batch file, because some ISP's only | give you a second or two between establishing a modem connection | and initiating PPP. You might also look at sub-pppd.bat for more | clues and ideas ... | | | | | ============================================================================ | | Fri, Apr 29, 2005 - 9:32:18 am | | gregy> | Ah...I can get miniterm to dialup and connect, then do the PPP packet | install, then it drops....BUT....Wizzy has something that allows you | to miniterm-connect/PPP packet install/then fire up Arachne as if you | had an ethernet/NIC always-on connection, or fireup Lynx, or use ping, | traceroute, etc...... | ... | That's what I was asking about, if LSPPP was capable of dialing, | connecting, installing PPPpacketdriver, then sitting there waiting for | you to do something with some internet app.... | .... | hmmm...where do I find sub-pppd.bat....??? I don't find it in any | of my Arachne installs (up to 1.83).... | | | ============================================================================ | | Fri, Apr 29, 2005 - 11:50:13 am | | phantom> | Hi gregy, | Š| Yep, I've used LSPPP as above, but not for quite a while. | Login to AD and check for the Lynx4AU packages, mentioned | on the main page. The let you switch between Arachne and | Lynx. | | | ============================================================================ | | Fri, Apr 29, 2005 - 12:01:55 pm | | phantom> | In other words, you can use Arachne as your main TCP/IP client, | including its dialling capabilities, exit whenever you like to | run other clients, such as Lynx, then return to Arachne to | continue or to finish your online session. | | | | ============================================================================ | | Fri, Apr 29, 2005 - 12:14:22 pm | | phantom> | In case you hadn't guessed, I prefer to do my dialling/connection | via Arachne, only resorting to more manual methods for diagnostic | purposes. | | Oh, from memory the Lynx4AU packages are in the Tools directory | in the AD Files area. If not, check the Links area for my web | site, also for the Lynx site. This stuff will give you an idea of | just how flexible this TCP/IP connectivity can be, using either | EPPPD or LSPPP. | | | | ============================================================================ | | Fri, Apr 29, 2005 - 12:24:40 pm | | phantom> | Final hint for the night ... If you've played with Miniterm and | LSPPP but get drop-out, I'd suggest starting from scratch with | a standard EPPPD configuration. Once that's working, it takes | just a few seconds with the LSPPPCFG package to convert to a | working LSPPP configuration. Then you'll be able to use Miniterm | and LSPPPCFG/LSPPP manually too, but you must first get it to | work via Arachne. BTW, the sub-pppd.bat file is generated when | you install the LSPPPCFG package. | | | ============================================================================ | | Fri, Apr 29, 2005 - 2:07:48 pm | | glennmcc> Š| No problem dialing/establishing PPP without Arachne. | | In arachne.cfg | | Connection Ready | | AT the DOS prompt while in the arachne_main dir | | lh lsppp.exe>ppp.log | | If lsppp.exe and lsppp.cfg are not in the arachne_main dir....... | | cd \lsppp | | lh lsppp.exe>\arachne_main\ppp.log | | There you go............. | | You are connected and ready to go when you decide to run arachne.bat | | | ============================================================================ | | Fri, Apr 29, 2005 - 2:10:07 pm | | glennmcc> | Joe, | | Thanks for the 'heads up' | | I just tried it and you are right... 0xb800 works perfectly in lsppp.cfg | | ============================================================================ | | Fri, Apr 29, 2005 - 5:51:23 pm | | ray> | Lots to think about. From what I hear, it's back to lspppcfg for me | since it maintains compatability with existing keywords, although I do | like the idea of an alternate lspppcfg that was built to include | dialer services as well for folks like me who don't have any | dialer issues with lsppp and would like to just dump miniterm. No | biggy though, as she is now, all is well. | | | | ============================================================================ | | Fri, Apr 29, 2005 - 5:54:41 pm | | glennmcc> | IIRC, | There was one particular release of lsppp.exe in-which the dialer | functions were 'broken'. | Š| As far as I know... everything is fine in all other releases. | | | | ============================================================================ | | Fri, Apr 29, 2005 - 6:29:02 pm | | ray> | Hmmm, then maybe there is nothing wrong with the current versions and | we can persuade Joe to update lspppcfg. If I followed the previous | remarks, the only remaining difficulty would be that one couldn't use | 'terminal' mode. That's never done anything for me although others | might have found it useful. | | ============================================================================ | | Fri, Apr 29, 2005 - 7:36:47 pm | | glennmcc> | terminal mode is only good for 'diagnostics' | | If you are having trouble logging on, you can follow the process | step-by-step to where it goes wrong. | | That is exactly the reason why I put this on pppglenn.htm which has now | been renamed as pppenhan.htm | | _ show terminal (diagnostic mode) | | ============================================================================ | | Fri, Apr 29, 2005 - 8:13:40 pm | | ray> | Ya, I know what it's for it's just a question of how useful it | still is. | | | ============================================================================ | | Fri, Apr 29, 2005 - 8:48:51 pm | | glennmcc> | It's _very_ useful every time there is a problem with login. | | ============================================================================ | | Fri, Apr 29, 2005 - 8:50:46 pm | | ray> | Well, then we still need the old keywords, no? | | | ============================================================================ Š| | Fri, Apr 29, 2005 - 9:02:15 pm | | glennmcc> | Some yes, some no. | | If you'll notice the list of keywords I felt could be removed....... | niether 'UseTerminal' nor 'TerminalWindow' is among them. | | | Port | Irq | Base | Mode | Speed | InitString | DialString | PhoneNumber | WaitFor1 | Response1 | WaitFor2 | Response2 | WaitFor3 | Response3 | PPPusername | PPPpassword | | [tcp/ip] | NameServer | AltNameServer | | | ============================================================================ | | Fri, Apr 29, 2005 - 9:17:25 pm | | glennmcc> | We don't need to use miniterm.exe as our terminal. | | I am logged-on right now using lsppp.exe as my 'terminal' | with these settings in arachne.cfg | | Connection @lh lsppp.exe /?>NUL | Dialer lsppp.exe /?>NUL | TerminalWindow @lh lsppp.exe | UseTerminal Yes | Autologin No | IP_Address BOOTP | | ============================================================================ | | Fri, Apr 29, 2005 - 9:32:58 pm | | glennmcc> | We will of-coarse need to use either miniterm or some other 'real' terminal Š| program to do a 'manual login' in which we enter our username and password | via the keyboard at the needed points during the login process. | | ============================================================================ | | Sat, Apr 30, 2005 - 7:08:57 am | | gregy> | Thanks, guys...for the comprehensive discussion... | .. | I think you've give me enough hints and tips to get me through the | trauma.... | | | | | ============================================================================ | | Sat, Apr 30, 2005 - 7:14:08 am | | gregy> | One of the things I wanted to do, was to just get an independent | LSPPP dialup/connect install working so I can check out Glenn's | report that his ISP actually gives him his DNSes (server-supplied | DNS) when he connects, just as we always knew our ISPs (most of them, | anyway) give us dynamically-assigned personal IPnumbers to use during | our online connection session. | | Miniterm/EPPPD cannot do that (or don't, anyway) and Glenn gave a | listing of his environment variables showing that LSPPP stashed two | DNSx numbers in there. If my ISP is doing this, also, that would | probably explain why the Shiva Dialer/Winsock could actually still | access DNSes even though I had reset its two DNSvariables to 0.0.0.0 | | | ============================================================================ | | Sat, Apr 30, 2005 - 7:52:21 am | | ray> | Glenn, | | But wouldn't the PhoneNumber still be needed to use miniterm? | Hell, it's the dialer no? | | | ============================================================================ | | Sat, Apr 30, 2005 - 2:36:39 pm | | glennmcc> | Yes, it would. | | But what I was talking about was elimiating miniterm. | Š| | | | ============================================================================ | | Sat, Apr 30, 2005 - 2:40:13 pm | | glennmcc> | All of these are only needed when using miniterm. | | Port | Irq | Base | Mode | Speed | InitString | DialString | PhoneNumber | WaitFor1 | Response1 | WaitFor2 | Response2 | WaitFor3 | Response3 | | | These are only needed when using epppd | | PPPusername | PPPpassword | NameServer | AltNameServer | | | When we use lsppp as both dialer and packet driver we no longer | need any of those lines in arachne.cfg | | All of that information is now in lsppp.cfg instead. | | | | ============================================================================ | | Sat, Apr 30, 2005 - 6:31:20 pm | | ray> | Yeah, I just meant if Miniterm was still needed for this terminal stuff. | | | ============================================================================ | | Sat, Apr 30, 2005 - 6:48:18 pm | | glennmcc> | It's not needed for 'terminal stuff'. Š| | As I showed... lsppp.exe can be used as the 'TerminalWindow' | | Miniterm (or an other terminal program we choose), | can be used for 'manual login'. | | But since almost all ISPs have now disabled the option of manual login | it is now a moot point. | | Virtually all ISPs now _require_ PAP/CHAP login and manual login will not wo | | | ============================================================================ | | Sat, Apr 30, 2005 - 6:52:56 pm | | glennmcc> | The only thing we need the terminal window for anymore is so that we can | see what's going on during dialup, connect, PAP/CHAP login, PPP negotiation. | | We only need to see this stuff when we are having a problem logging in. | | By choosing to use the terminal window, | we can diagnose what the problem is. | | That's why I marked it (diagnostics mode) on pppenhan.htm | | ============================================================================ | | Sat, Apr 30, 2005 - 8:48:30 pm | | ray> | Ok, got it ... more or less ;-) ... but if actual manual login | is obsolete, then the terminal window is just a viewer, no? | And if that's true, can't its output be logged to a file for viewing? | If so, then old miniterm would truely be history, no? | | BTW a bit of a milestone here: core.exe 397,856 | | | ============================================================================ | | Sat, Apr 30, 2005 - 8:51:59 pm | | ray> | Oh, almost forgot, you don't suppose that problem of Bastiann's is | due to your cutting back on static modules do you? If I was you, I'd | send him a core with the modules as before, and see if that fixes his | problem since that could prove to be an important bit of information | as to what can and can't be overlayed. And we both want to fart | around with that. | | | | Š| ============================================================================ | | Sat, Apr 30, 2005 - 9:55:10 pm | | glennmcc> | It already _is_ logged to a file. | | That's the reason for lsppp.exe>ppp.log | | | | ============================================================================ | | Sat, Apr 30, 2005 - 10:05:25 pm | | ray> | So then miniterm really is rather marginal, no? | | | ============================================================================ | | Sun, May 01, 2005 - 2:01:40 pm | | glennmcc> | yes.... 'marginal'. | | Maybe we still need it..... maybe not. | | | ============================================================================ | | Sun, May 01, 2005 - 8:07:25 pm | | glennmcc> | I've been doing some experimenting this evening and I just did a | successful fresh compile of v1.66b ......... | | | are you ready for this ? | | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - Š| - | - | - | - | - | - | - | - | - | - | - | - | - | | for Linux svgalib :) | | ============================================================================ | | Sun, May 01, 2005 - 9:42:14 pm | | ray> | He he, did you really get all the graphics calls migrated? Sounds too | good to be true, but if it is as you say that's f***ing marvelous, | pardon my Canadian French. We now have a doorway to the future. | | | ============================================================================ | | Mon, May 02, 2005 - 12:11:01 am | | Joe> | Glenn, | | I think Port, Irq, Base, Speed may still be required without Miniterm. | (Perhaps Speed is optional. IIRC, one either uses Port or Base/Irq, | not both.) | | | ============================================================================ | | Mon, May 02, 2005 - 12:17:16 am | | Joe> | Gregy, | | Dynamic DNS is standard fare these days and has been for quite a | while. It's been supported by LSPPP for quite a while too, and | IIRC, has always been the default configuration with LSPPPCFG | (since ca. 2000). In fact, a number of ISP's in Oz will not allow | you to select a different DNS, only whatever their server tells | your TCP/IP application. The way it works is that 0.0.0.0 is not | a valid IP, so the server says to the application, "No, use these | IP addresses for DNS instead". The ISP's I use here will say that | for non-zero IP addresses too. | Š| | ============================================================================ | | Mon, May 02, 2005 - 9:36:55 am | | glennmcc> | Hi Joe, | | Nope.... not needed in arachne.cfg _if_ miniterm/epppd are not used. | | When we use lsppp as both dialer and packet driver, all of the needed | information is in lsppp.cfg instead of arachne.cfg | | ---- my current settings in arachne.cfg --- | | [profile] | Profile 4234005L.ACF | Connection @lsppp.exe /?>NUL | Hangup @termin.com 0x60>NUL | | [dialer] | DialPage file:pppenhan.htm | | Dialer @lh lsppp.exe>ppp.log | UseTerminal No | Autologin No | | [tcp/ip] | IP_Address BOOTP | IP_Grab IP address set to | Gateway 0.0.0.0 | AltGateway 0.0.0.0 | Netmask 0.0.0.0 | TCPconfig PPPTCP.CFG | ___________________________________________ | | | --- my current settings in lsppp.cfg --- | /M:atf50 | /x | /i:4 | /b:bx800 | /d:330-423-4005 | /B:115200 | /V:60 | /U:glennmcc@cisnet.com | /P:***** | ________________________________________ | | As you see... | all of the lines I mentioned have been deleted from arachne.cfg | and the ones that are needed for lsppp are over in lsppp.cfg instead. | | ============================================================================ | Š| Mon, May 02, 2005 - 9:46:40 am | | glennmcc> | When using miniterm to dial _and_ a special base port address is needed. | (such as in my case of 0xb800) | | _both_ port # _and_ base port address must be used. | | This is due to the fact that 'port 0' is the signal to miniterm.exe to | use 'base #' instead. | | If we delete the port line (or comment it out with ; preceeding it), | miniterm will not dial because it uses 'port 2' by default when | 'port #' is missing from arachne.cfg | | | ============================================================================ | | Mon, May 02, 2005 - 8:15:12 pm | | ray> | I done a good thing today: I've created the keyword "Experimental" | which let's the user turn off anything experimental with one keyword. | This permits me have a 'stable' core and an 'experimental' core | at the same time. I'd give it some thought Glenn, it would let | your customers back out of anything iffie and thus allow you to | experiment more without having to worry about new bugs poping up | and having to scramble to patch it. | | | ============================================================================ | | Mon, May 02, 2005 - 9:11:52 pm | | glennmcc> | I prefer this method..... | | #ifndef GLENNMCC | //experimental stuff | #endif | | | This way the experimental stuff does not get compiled into the release | version but only into the ones I compile for myself using glennmcc.prj | | | | ============================================================================ | | Mon, May 02, 2005 - 9:13:19 pm | | glennmcc> | Oops, typo... | | That actually reads "#ifdef GLENNMCC" not "#ifndef GLENNMCC" Š| | | ============================================================================ | | Mon, May 02, 2005 - 9:17:16 pm | | glennmcc> | Since I always use nokey.prj for the release versions. | This method also works great. | | #ifndef NOKEY | //experimental stuff | #endif | | ============================================================================ | | Mon, May 02, 2005 - 9:53:48 pm | | ray> | Yeah, that's cool too, but it serves a slightly different need. | I'd say my method is more for code that is considered ready for | prime time, but the config option would be used as a failsafe in case | of some unforseen problem, or just in case some new feature wasn't | liked. I'd use the conditional for very raw experimental stuff | or for some custom build or other. | | BTW any word from Bastiaan on that stutter? I can't remember if I | mentioned it or not, but I'm running right now with an overlay | buffer forced to 128K, and the improvement in speed is quite nice. | It will be interesting to compare that with various modifications | to the number of modules made static. | | | ============================================================================ | | Tue, May 03, 2005 - 6:55:52 am | | phantom> | Hi Glenn, | | Yes, you are correct about Port being required in any case, if only | to indicate "use Base and Irq instead", ie. the special value 0. I | haven't tried simply omitting it, perhaps this will have the same | effect? | | Now as for all those keywords being "redundant" without Miniterm, | the problem here is a difference in perspective. I consider it | Arachne's job to control its dialler and packet driver, in which | case this is either done directly (for Miniterm/EPPPD) or indirectly | (for Miniterm/LSPPPCFG/LSPPP). If we add lsppp.cfg support to Arachne, | then once again, these keywords will be used directly to control | LSPPP. I prefer all the configuration screens and ACF's and wizards | to work in configuring these settings, not a separate configuration | file (ie lsppp.cfg) that the user must manually edit if they want to | use LSPPP. One of my primary goals in writing LSPPPCFG was to eliminate Š| the need to edit any file whatsoever, to use LSPPP with Arachne in | exactly the same manner as for EPPPD. | | | | ============================================================================ | | Tue, May 03, 2005 - 5:22:29 pm | | glennmcc> | re: | "haven't tried simply omitting it, perhaps this will have the same effect?" | | Nope.... omitting it causes miniterm to default to 'port 2' | | As for the rest......... You are 100% correct. | | If you can modify lspppcfg.exe to 'grab' all of those lines from | arachne.cfg and place them into lsppp.cfg in thier correct formats. | | We will then have 'perfection'. ;-) | | | | ============================================================================ | | Wed, May 04, 2005 - 12:41:24 am | | Joe> | Hi Glenn, | | Port 2, eh? | | Yuck!!! | | ============================================================================ | | Wed, May 04, 2005 - 3:48:54 pm | | glennmcc> | Yep..... Yuck is right. | | This is just one more reason to DITCH miniterm and go with lsppp for | _both_ dialer and packet driver. | | | ============================================================================ | | Wed, May 04, 2005 - 5:09:23 pm | | ray> | Yup, let's stick with one proven product and make it work. | | | ============================================================================ Š| | Wed, May 04, 2005 - 5:56:20 pm | | ray> | What is 'chk-optn' as mentioned in 'sub-pppd.bat'? I can't find | that as a command or even as a string anywhere. | | | ============================================================================ | | Thu, May 05, 2005 - 6:00:26 pm | | Joe> | Ray, | | 'chk-optn' is a utility from the LSPPPCFG package which is used | to check "option settings" in '$roura$.bat'. It is used to determine | what command line option the LSPPPCFG package should use to invoke | 'arachne.bat'. This feature was required because your Arachne | configuration fetched e-mail first, rather that going to a default | (home) web page, early versions of the LSPPPCFG package assumed the | latter and always used the "-o" command line option. | | | | ============================================================================ | | Thu, May 05, 2005 - 8:50:50 pm | | ray> | Thanks Joe, | | I found it using a wider search, I had managed to loose it in my | working copy of Arachne, so I was running without if for some time | without any noticable problems, but as you say, this seems to | be a responce to a specific situation. | | | ============================================================================ | | Sat, May 07, 2005 - 8:25:38 am | | ray> | What's the option that makes the arrow keys move the cursor instead | of scrolling the document? | | ============================================================================ | | Sat, May 07, 2005 - 10:54:03 am | | ray> | Joe, Glenn, | | I'm having fun butchering the dialer code, got it cut down to | about 1/3 the original size. There will be mistakes in this, but Š| it's been instructive. A question: what are the errorlevel values | returned by lsppp.exe, what, when and why? | | | | ============================================================================ | | Sat, May 07, 2005 - 2:45:31 pm | | glennmcc> | Realtime smooth scrolling ("Netscape") | | ============================================================================ | | Sat, May 07, 2005 - 8:25:03 pm | | Joe> | Me confused! | | I thought the answer was : the shift key! | | ============================================================================ | | Sat, May 07, 2005 - 8:28:54 pm | | Joe> | Ray, | | Well, early versions of LSPPP did not set the errorlevel | on exit - it was random (in the sense of being meaningless). | | On my suggestion (insistance;-), DL fixed this, however I'm | not sure without studying the code if there's specific levels | used, or if it's just a matter or zero vs. non-zero. | | | ============================================================================ | | Sat, May 07, 2005 - 8:31:47 pm | | Joe> | Typo ... that should have read as : | ~~~ a matter of zero vs. non-zero. | | ============================================================================ | | Sat, May 07, 2005 - 10:55:11 pm | | glennmcc> | Hi Joe, | | You not confused..... shift is method #2 ;-) | | Turning off Smooth scrolling eliminates the need for holding the shift | key to get the arrow keys to move the mouse pointer. Š| | ============================================================================ | | Sun, May 08, 2005 - 6:10:48 pm | | ray> | Glenn, Tx, I always forget that one. BTW, why doesn't some clever | hotshot combine the two so that if you are using the cursor movement | setting, once the cursor hits the bottom of the screen the page starts | scrolling? This would be a third option for 'Smoothscroll' I figure. | | Joe, | | The reason I ask is that I'm trying to figure out this, which is | the contents of $roura$.bat when the dialer is invoked: | | | /* | @lh lsppp.exe /d:6046788350 ..... | @if errorlevel 1 goto skip | @lsppp.exe /?>NUL | :skip | @if exist IP-UP.BAT call IP-UP.BAT | @echo PPPD status: IP address set to %MYIP%>>PPP.LOG | @arachne.bat -r | */ | | The way it looks, an errorlevel of one is OK and it skips the | next line which as far as I can tell does nothing useful anyway. | I've butchered all of that out anyway to see what happens and it seems | fine. Next I'll try to incorporate what I think your lspppcfg does | inside the dialer and see what I can come up with. | | | | ============================================================================ | | Sun, May 08, 2005 - 6:20:49 pm | | ray> | Oh, Glenn, I noticed that rotten old smartdrv solved B's problem | but I'd like to know which file is the culprit, ie the file that | needs to be reloaded so often, that would be a good candidate for | static link no? I'm all the more curious since when you sent him | a build with the original plan, it didn't help. Perhaps there is some | file that never was static that should be. | | | ============================================================================ | | Sun, May 08, 2005 - 6:47:41 pm | | glennmcc> | It's in the Internal Editor stuff.... | Š| The IE*.c files. | | BRB | | | ============================================================================ | | Sun, May 08, 2005 - 7:54:14 pm | | ray> | Hmmm, the obvious suspects. | | | ============================================================================ | | Sun, May 08, 2005 - 8:19:54 pm | | glennmcc> | Nope, I was wrong.... that was not it. | | But then I can't duplicate Bastiaan's "stuttering" problem using any of | the release versions of core.exe | | ============================================================================ | | Sun, May 08, 2005 - 9:23:57 pm | | ray> | Could be some one-off anomaly not worth chasing or maybe the | speed of his machine is a factor -- IIRC it was rather slower than yours | but then there are several people running relatively slow machines | with no trouble. | | ============================================================================ | | Mon, May 09, 2005 - 10:46:14 am | | gregy> | Joe....thanks for the confirmation. My ISP swears they don't | auto-supply the DNSnumbers...but...the operation of that Win3.1 Shiva | Dialer says they do......that is why I want to investigate this using | LSPPP. | ... | However....please do not refer to this as "Dynamic DNS"...if you google | that, you will find that it refers to something very different that | what we are talking about. | ... | Google "Server supplied(provided) DNS" (as Glenn did) and you will | come up with what we are talking about. This "INET Terminology" problem | has bit us a couple times in the past (that I can remember)....people | talking about one thing, while using terminology that refers to quite | another thing. | | That's what thru me off, and perhaps thru my ISP's tech people off (I | havent' gone back to them with the term "server supplied DNS"..so maybe Š| they're not so clueless). But, those two different terms are what | threw me off, and got me going in the wrong direction. | ... | he he...thank GOD we gots lotsa "redirectors" around here... | | | | ============================================================================ | | Mon, May 09, 2005 - 10:47:44 am | | gregy> | Glenn, Ray.....I could not reproduce Bastiaan's stutter using 1.83 on | this Cyrix 486dlc 40MHz box, so I don't think it's due to chip speed. | .. | | | ============================================================================ | | Mon, May 09, 2005 - 2:50:04 pm | | ray> | Gregy, as one last test, try running it with your cache disabled, if | that doesn't reproduce the problem then it would seem to be some | local anomaly of Bastiann's | | | ============================================================================ | | Mon, May 09, 2005 - 4:59:33 pm | | glennmcc> | I am on line right now with v1.83 core.exe running 100% on diskette. | No diskcache loaded at-all. | stutter...stutter... stutter... | BRB with the compile I made for bastiaan that has all of modules back | the way they originally were. (static vs overlayed) | | | ============================================================================ | | Mon, May 09, 2005 - 5:11:48 pm | | glennmcc> | Now I'm back with the core.exe I compiled on May 1st | Still stuttering :( | BRB | | | ============================================================================ | | Mon, May 09, 2005 - 5:21:27 pm | | glennmcc> | Now let's see ............ Š| | Here's a string a of random characters typed very quickly to see what happen | | .kjndvc:SJHF;OIENGF/AKFLJHT90[2QU536QPAKINBVZ[9YUHq]nBslg'dap]t90-qjh32yt' | ;oasmnbvaxjlc]opjtgA;MLGZx:MLBVASKTRP.,mvn:LKXNVCKLNfsKLANFDLKanf/;KLznFk | lansfd/;KLANF/;klansfkl;ANSF/;KLandflkNADFklandklfANSDFKL;anfl;kasndgfkla | nsklgnasklgfdnadsklgna/;lskbvnasklngd | | We got it...... NO STUTTERING AT-ALL !!! :))) | | ============================================================================ | | Mon, May 09, 2005 - 5:24:40 pm | | glennmcc> | Comparing my files I find that I screwed-up back on May 1st and still | had the .PRJ in-place which had some additional modules overlayed. | | This compile I just made now has everything back the way it was. | | | ============================================================================ | | Mon, May 09, 2005 - 5:27:59 pm | | glennmcc> | I will upload a ZIP to my site in a few minutes with everything that's | on this diskette. | | This sucker is a perfect way of testing disk activity. | | ============================================================================ | | Mon, May 09, 2005 - 5:45:13 pm | URL | | glennmcc> | Unzip to A:\ with........ | | pkunzip -d diskette.zip a:\ | | Also place your own ARACHNE.CFG, ARACHNE.PCK and LSPPP.CFG | over onto the diskette. | | Also copy MINITERM.EXE and EPPPD.EXE onto the diskette | if you use them for dialer/packet driver instead of LSPPP.EXE | | Alter the path lines in ARACHNE.CFG to point to A: as-well. | (make new dirs as needed on A:\) | | ============================================================================ | | Mon, May 09, 2005 - 6:19:42 pm | | ray> Š| Cool, all this is grist for the mill as far as understanding goes. | Please advise when you find the exact culprit. | | | | ============================================================================ | | Mon, May 09, 2005 - 9:32:32 pm | | gregy> | ummm...Ray...I have to report a "local anomalie" of my own.... | ... | I doan hav no steengkin diskcache........In this DRDOS 6.0 it | is PCKWIK and I have all reference to it removed from my boot files. | ... | I've had it this way ever since I figured out that the ONLY way to | run Arachne is from a RAMDisk.... | .... | | | | ============================================================================ | | Tue, May 10, 2005 - 3:14:03 am | | Joe> | Glenn, | | Thanks for the info about cursor keys and Smooth Scrolling, | I had *no* idea they were in any way related! | | ============================================================================ | | Tue, May 10, 2005 - 3:16:24 am | | Joe> | Ray, | | Lots of fun with your experimentation. FYI, I don't think LSPPPCFG | in its present version is quite what you need. I have been giving | some thought to how I should enhance it for LSPPP dialling, I'll | look into that after doing the "clean code" version of 1.83J1 ... | | | ============================================================================ | | Tue, May 10, 2005 - 3:27:39 am | | Joe> | Gregy, | | Thanks for the "heads up" on the "Dynamic DNS" jargon. It seems they | (who?) want to (re?)define this as "DNS support for Dynamic IP", so | no wonder there's some confusion here. Silly people, huh? | Š| ============================================================================ | | Tue, May 10, 2005 - 1:31:29 pm | | ray> | Gregy, | | Whaaa, are you a wise-guy or sometheeen? ;-) | | I forgot you're a ramdisker. I suppose that would have more or less | the same effect on speed issues as would cacheing, or better, so | there's little to be learned there. | | | ============================================================================ | | Tue, May 10, 2005 - 2:04:13 pm | | ray> | Joe, | | Roger that. FWIW, I'll post my boiled down dialer code on AD. | | | | ============================================================================ | | Tue, May 10, 2005 - 6:28:18 pm | | ray> | What's our established minimum free memory? I've been running for | some time with my core set to be rather glutonous so as to leave only | about 106K free, and it seemed ok until I ran a script and crashed. | | | ============================================================================ | | Tue, May 10, 2005 - 7:28:13 pm | | glennmcc> | IMO, | | 450 is the absolute min (_IF_ one is able to load the packet driver high) | 500 to 550 when conv mem must be used for the packet driver. | | ============================================================================ | | Tue, May 10, 2005 - 7:32:37 pm | | glennmcc> | My current stats (lsppp loaded high) | | DOS mem (bytes) 200672 | | --- with core.exe 'dumped' (memc.dgi) --- Š| | DOS 91,280 ( 89K ) 16490 | QEMM386$ 768 ( 0K ) 300 | COMMAND 1,936 ( 1K ) 790 | FREE 642,928 ( 627K ) 9CF70 | Largest executable program: 642,912 ( 628K ) | Total Free DOS memory: 720,864 ( 704K ) | | ---- with core.exe still in memory (memcore.dgi) --- | | DOS 91,280 ( 89K ) 16490 | QEMM386$ 768 ( 0K ) 300 | COMMAND 1,936 ( 1K ) 790 | CORE 442,768 ( 432K ) 6C190 | command 7,200 ( 7K ) 1C20 | FREE 192,960 ( 188K ) 2F1C0 | Largest executable program: 192,944 ( 188K ) | Total Free DOS memory: 270,896 ( 265K ) | | | | ============================================================================ | | Tue, May 10, 2005 - 7:37:02 pm | | glennmcc> | by "ran a script"... do you mean an .ASF ? | | If so... | I have found that .ASFs containing 'reload:' usually require no less than | DOS mem 150000 to prevent a crash. | | IMO, | .ASF handling needs a 'major overhaul'. | | ============================================================================ | | Tue, May 10, 2005 - 9:22:35 pm | | ray> | Yes to all. I think the magic number might be 128K free (core.exe loaded | of course) I did some crude memory destroying tests and found that | both our cores are ok with 'reloading' ASF's if 128K+ is available. | That's the size of two 64K malloc() buffers so it seems likely. | So what is it about ASF's that require such huge buffers? If I can | find a more accurate way of filling up my memory, I'll run more | exact tests. | | | ============================================================================ | | Thu, May 12, 2005 - 9:22:33 pm | | ray> | Fixed a wee bug that's been bothering me for some time: Š| You know how in a 'compose mail' screen, if you move the mouse | directly down from the input area onto a button if forgets it's color? | I think I've bagged it. | | And, in the process I've added a third mouse color for when it's on | a text area. | | | | ============================================================================ | | Thu, May 12, 2005 - 10:57:55 pm | | ray> | So Glenn, what about those colours? What's the pattern there? | | | ============================================================================ | | Fri, May 13, 2005 - 5:01:30 pm | | glennmcc> | ????? | | I don't understand your question. | | ============================================================================ | | Fri, May 13, 2005 - 5:56:53 pm | | ray> | If you set your mouse colors to '2 2' your cursor will turn RED. | If you use '3 3' you get GREEN ... this is hard to explain using | either IRGB or RGB or RRGGBB or anything else that I can think of. | It doesn't matter a whole lot of course, one can simply take it on | faith that 3 == GREEN, but I'd like to know why. | | | ============================================================================ | | Fri, May 13, 2005 - 8:19:26 pm | | glennmcc> | Here's what I've just found by checking 0 to 15 in MouseColors | | 0 black | 1 dark blue | 2 dark red | 3 dark green | 4 dark purple | 5 dark aqua | 6 dark yello | 7 light gray | 8 dark gray | 9 bright blue Š| 10 bright red | 11 bright green | 12 bright purple | 13 bright aqua | 14 bright yellow | 15 white | | | Checking into config.c I find that the CHAR value of MouseColors is | the int of MouseColors after being converted into a longint | | (int) strtol(value, n10) | | MouseColors 0 3 == the int of the longint 03 in base 10 | | I think that there MUST be a better way. | | We are getting 16 color possibilities..... but theye don't match | the 'standard 16 color table' | | 0 BLACK | 1 BLUE | 2 GREEN | 3 CYAN | 4 RED | 5 MAGENTA | 6 BROWN | 7 LIGHTGRAY | 8 DARKGRAY | 9 LIGHTBLUE | 10 LIGHTGREEN | 11 LIGHTCYAN | 12 LIGHTRED | 13 LIGHTMAGENTA | 14 YELLOW | 15 WHITE | | | ============================================================================ | | Fri, May 13, 2005 - 8:21:04 pm | | glennmcc> | Gregy !!!!!!!! | | Fix this, will ya please ? | | (int) strtol(value, & newvalue, 10) | | ============================================================================ | | Fri, May 13, 2005 - 9:19:50 pm | | ray> | Yes, that's my point, the lopif functions have their own strange Š| ideas about colors. I used to be just vaguely disturbed by it | and now I know why, since one would expect the standard values to | be used. BTW, the strtol isn't the problem since I long ago | started using a simpler parsing routine for those values, you can | plug numbers directly into the functions and it makes no difference. | | | | | ============================================================================ | | Fri, May 13, 2005 - 9:21:39 pm | | ray> | What puzzles me even more is that exactly half of the results we | see *do* match the standard table -- '14' is bright yellow just | as one would expect it to be. Puzzling, no? | | | ============================================================================ | | Fri, May 13, 2005 - 10:45:16 pm | | glennmcc> | Ah HA !!! | | So, the broblem is not in the SRC file(s) of core.exe | | It's somewhere in the SRC files of HIXLOV.LIB | | ============================================================================ | | Sat, May 14, 2005 - 8:31:05 am | | ray> | Exactly, this ain't our problem, it's Haro's problem. Without Michal, | it's hard to go further except to just be aware of the discrepancy. | | | | ============================================================================ | | Sat, May 14, 2005 - 2:53:31 pm | | glennmcc> | OK, sounds good to me. | | Suggestion.... | | How about if we include that color table in the documentation ? | | BTW, | It applies to HTML BGcolor and text color as-well. | | Arachne.cfg variable Š| | Colors n n | | Works in exaclty the same manner as does | | MouseColors n n | | Therefore this color table applies to both of those CFGvariables | | 0 black | 1 dark blue | 2 dark red | 3 dark green | 4 dark purple | 5 dark aqua | 6 dark yello | 7 light gray | 8 dark gray | 9 bright blue | 10 bright red | 11 bright green | 12 bright purple | 13 bright aqua | 14 bright yellow | 15 white | | | ============================================================================ | | Sat, May 14, 2005 - 3:02:26 pm | | glennmcc> | Oops... my bad. :( | | Colors n n | | Does not refer to HTML BGcolor and text color | | It refers to the text color and BGcolor of text input fields | | Colors 0 15 | | Makes this text area, the input fields for Nick:, Email:, and URL: | and the URL bar showing the current page URL .... black text on white BG | | ============================================================================ | | Sat, May 14, 2005 - 7:00:27 pm | | ray> | Yup some documentation would be apropriate, we both have the facts | now. BTW, the affected functions AFAICT are: x_setcolor(), x_setfill() | and x_yncurs() (names from memory!) there may be others but I haven't | found them. I now have defines in place for all calls (which is what | I set out to do several day ago ;-) Š| | BTW I have a bug report: Go to the move mail page, pull down the | menu for other folders and move the mouse rapidly off the menu and | back on to neutral territory, what happens? | | | | ============================================================================ | | Sat, May 14, 2005 - 9:12:41 pm | | ray> | You got no problems there? | | | ============================================================================ | | Sat, May 14, 2005 - 11:20:49 pm | | gregy> | Glenn, I will try to put the new-style board up this next week... | ... | I am sorry that I have been so sorry..........this winter I | have not been able to do or complete very many things... | | | ============================================================================ | | Sun, May 15, 2005 - 8:39:39 am | | ray> | Ha, well spring planting can't be far off and you won't have any time | then ;-) | | | | ============================================================================ | | Sun, May 15, 2005 - 2:49:11 pm | | glennmcc> | Thanks gregy. | | BTW, | | Please be sure to 'archive' this board before making the changes. | | ============================================================================ | | Mon, May 16, 2005 - 10:34:35 pm | | gregy> | yes...I'll d/l the whole thing, convert to print, then split it up | by quarter-years, or something... | .. Š| | | ============================================================================ | | Tue, May 17, 2005 - 9:45:53 pm | | ray> | Glenn, | | You'll have to kill me. Downloaded 1.83, poached your fix for that | SELECT thing and she works great. Inspired by your insights, I | went a little further and managed to simplify the code quite a bit. | Still, it's just a hack; why should the SELECT tag cause an endless | loop? OnMouse() should be able to handle that, or, second best, | the specific call to MouseOn() that's causing the fault should have | the filter aplied there. | | | | ============================================================================ | | Wed, May 18, 2005 - 12:10:21 pm | | glennmcc> | Hey, hey, hey, watch it there bud !!! | | That's _my_ hack you're jabbing at. | | Yes, it's a hack... and I'm damned proud to be a hacker. | | ============================================================================ | | Wed, May 18, 2005 - 1:03:23 pm | | ray> | Sure, it's a hell of a lot better than mine was but still a hack. | All I'm saying is we should keep on the lookout for a fix that's | closer to the heart of the problem. | | | ============================================================================ | | Wed, May 18, 2005 - 4:40:22 pm | | glennmcc> | Now he wants to stab me in the heart. | | Sheesh...... sometimes a guy just can't win. | | | | | ============================================================================ | | Wed, May 18, 2005 - 4:43:47 pm Š| | glennmcc> | But seriously now... | | If it ain't broke don't fix it. | | So if the fix ain't broke don't fix it. | | ============================================================================ | | Wed, May 18, 2005 - 4:45:42 pm | | glennmcc> | If you feel like fixing something that REALLY needs fixing..... IgnoreJS | | | | ============================================================================ | | Wed, May 18, 2005 - 5:47:20 pm | | Joe> | Hi Ray, Glenn, | | What particular piece of "select" code are you two talking about??? | | | ============================================================================ | | Wed, May 18, 2005 - 7:27:42 pm | | ray> | Joe, | | It's the bug fix for the crash that happens (happened!) if you | move the mouse off a pulldown menu. Seems the HTML tag involved | is called 'SELECT' ... but what do I know. See Glenn's fix in | MouseOn(), mouseon() in the old code, ver 1.83 or better. | | BTW, Joe, I now have an Australian Nephew in law, so I guess we're | related ;-) Can you give me a few choice Ausie insults that I might | use on him? The kidding sort, for teasing him. | | | ============================================================================ | | Wed, May 18, 2005 - 7:31:21 pm | | glennmcc> | Hi Joe, | | This be the code in question :) | | --- in guitick.c ---- | Š| --- near the top --- | | #include "html.h"//!!glennmcc: Feb 18, 2005 -- for 'Select test' in mouseon( | | --- further down in the mouseon() function itself --- | | void mouseon(void) | { | if(!global_nomouse) | //!!glennmcc: Feb 15, 2005 -- fix pointer colors and 'stuck pointer' | { | //!!glennmcc: added 'Select test' Feb 18, 2005 | //calling onmouse(0) while in a | x_yncurs(1,mousex,mousey,15); | } | //!!glennmcc: end | } | | _____________ | | ============================================================================ | | Wed, May 18, 2005 - 7:32:33 pm | | glennmcc> | | | Hi Ray :) | | | | | ============================================================================ | | Wed, May 18, 2005 - 8:43:38 pm | | ray> | Oh, I don't think mine will work in your code, basicaly I've moved | all mouse color work into onmouse() itself so MouseOn does nothing | but filter the SELECT tag, as you figured out, then it calls OnMouse(), | that's all. In fact, if I can figure out where this bad SELECT | problem comes from, I'll remove MouseOn entirely and just call | OnMouse directly. Besides MouseOn relies on the return value of | OnMouse to set one of two colors, but right now I'm using four colors: | Link, Neutral, Text, and Menu -- yes I know, it's just BS but | I'm having fun. If you're intested I'll post my OnMouse(). Š| | | ============================================================================ | | Thu, May 19, 2005 - 1:48:35 am | | glennmcc> | DUDE......... Of course I'm interested. :) | | As for having fun....... I'm with you on that one. | | Why else are we doing this if not to have fun ? | | ============================================================================ | | Thu, May 19, 2005 - 5:21:34 pm | | ray> | OK, it's hardly finished work but I'll upload it to AD, I think | it's adaptable to the old code, just look for the new macro MOUSE. | | | | ============================================================================ | | Thu, May 19, 2005 - 8:25:19 pm | | ray> | Ha, found the offending calls: If you want, you can move your filter | from MouseOn() into HideHighlight() and ShowHighlight() ... just filter | the call to MouseOn(). That's not your style, but it will be faster, | since the filter is only checked when needed. | | | ============================================================================ | | Fri, May 20, 2005 - 9:49:20 pm | | ray> | Think I've got the *real* fix, or getting very close anyway -- no | need for filters. | | | ============================================================================ | | Sat, May 21, 2005 - 5:23:34 am | | Joe> | Hi Glenn, | | Thanks for the code above ... Yes, it does look like a slight | kludge, but nothing worth worrying about. OTOH, tracking down | the real culprit is hours of fun and helps to gain knowledge | of the code, right Ray? | Š| | | ============================================================================ | | Sat, May 21, 2005 - 5:37:11 am | | Joe> | Hi Ray, | | Thanks for the uploaded code at AD, I'll study it a bit later (my | 'puter has been having lock-up problems these past couple of days, | just found and fixed the problem - an electo in the power supply | was starting to "go"). | | BTW, you might like to post an explanatory note at AD to accompany | your posting, there's 33 subscribers there, but I don't think many | of them read the "Board" (it doesn't forward them any postings by | e-mail;-). Also, if I can make a request, it would be best to try | to remember to include a Subject heading with all postings, it makes | threads easier to follow or to look up in the future (TIA:-). | | | ============================================================================ | | Sat, May 21, 2005 - 5:54:07 am | | Joe> | [OT] | | Ray, | | I've been thinking about "insults" for you (for your nephew-in-law). | I'm probably not a good source for these, since I don't generally | engage in this practice (sure, I kid around with friends I know | well, but it's very specifically targeted and doesn't really involve | "insults" in any case). However, some guidelines : | | Firstly, don't start trying such "insults" until you know he won't | take them seriously! You may refer to him as a Kiwi; imply he's a | convict or of convict stock; if he's from NSW, refer to him as a | Victorian (or vice versa); find out what football team he follows | and kid him when they lose or their arch-rival is doing well. Hope | that's some help. If you have a particular "insult" in mind and | want an australian flavour to it, just ask (as long as you don't | blame me if it all goes wrong!;-). | | | ============================================================================ | | Sat, May 21, 2005 - 3:13:19 pm | | glennmcc> | Cludge ???? | | What make THIS a cludge ???? Š| | #include "html.h" | | | | if(activeatom.data1!=SELECT) | | | | else | x_yncurs(1,mousex,mousey,15); | | | Thats the whole fix..... 4 lines and nothing more. | | | ============================================================================ | | Sat, May 21, 2005 - 3:33:23 pm | | ray> | Joe, | | Roger all that. I'll try the 'state' thing, he's from Sydney so | that's Victoria IIRC so I'll call him a 'bloody Queenslander' or | something, thanks :-) | | Glenn, | | Tisk tisk tisk, you don't know a kludge when you see one? | Sure it fixes it, but at the cost of a filter that is used | 100's of times when it isn't needed just to catch one as yet | undiscovered problem. I've almost found the exact cause of the | problem. And the right fix looks like it will remove lines, not | add them. | | | ============================================================================ | | Sat, May 21, 2005 - 7:45:02 pm | | ray> | Well, it looks good. Code to AD tommorow. | | | ============================================================================ | | Sun, May 22, 2005 - 1:45:10 am | | gregy> | Joe, Ray, | | If Ray was from US (instead of that strange land just to the north), | since his nephew-in-law had married into a yankee family...he could | call him a "bloody septic tank"...and insult him royaly....but...since Š| Ray's a Canuck, heck....that's no option...... | | | | ============================================================================ | | Sun, May 22, 2005 - 10:30:20 am | | ray> | Ha, I love that, that insult should be universalized. | | ============================================================================ | | Sun, May 22, 2005 - 11:29:43 pm | | Joe> | Hi Glenn, | | Hey, don't fret ... a simple kludge beats an unresolved | problem any day of the week or year!!! :-) | | As I said before, it's nothing to worry about. Ray seems | keen to track this one down to its root cause, so lets see | what turns up ... | | | ============================================================================ | | Sun, May 22, 2005 - 11:33:36 pm | | Joe> | [OT] | | Very funny, Ray! | | | ============================================================================ | | Sun, May 22, 2005 - 11:51:00 pm | | ray> | Oh ya, I should post that. To be brief, the ultimate cause is | in ... (nuts can't remember the func name) ... SelectSwitch() ?? | ... where there is no guard against running off the menu downwards. | Basicaly, SelectSwitch calls MouseOn, calls OnMouse, which tries to | draw a highlight where there is nothing to highlight which then | returns a bogus value back to SelectSwitch which ends up in a loop. | | You can break the cycle in MouseOn, but that results in the mouse | colors misbehaving (check 1.83) so I put a guard right in SelectSwitch | which notifies OnMouse not to try to highlight a nonexistant button. | | Clear as mud? I'll post it manana. | | Š| ============================================================================ | | Sun, May 22, 2005 - 11:55:17 pm | | ray> | BTW, Joe, I wrote some vastly simpler dialer code that works fine for | me, (I figure I've sort of built in lspppcfg right into core). | It cuts out minitrue altogether, and just uses lsppp for everything. | Problem is it doesn't work for Vladimir. If he uses lsppp as his | dialer, he can't go on to ppp. Do you have any idea why that might | be? I know you mentioned that there were problems with lsppp as | a dialer, maybe this proves that that problem is still with us. | Comments? | | ============================================================================ | | Mon, May 23, 2005 - 1:37:36 am | | Joe> | Hi Ray, | | Miniterm, Miniture, it's all the same, no? ;-) | | Anyway, re LSPPP dialling - IIRC it works for me, although | it's been a looong time since I've tried and that was probably | with a defunct ISP anyway. It's one or two other users that | have complained about it (being unable to dial and connect | using just LSPPP) to me over the years, which has made me | stick with Michael's Miniterm. I am currently migrating to a | new PC at work, once that's done and things are back to normal, | I'll see if I've archived any of those e-mails about this issue. | Since I couldn't reproduce the problem at the time, it was not | practical for me to try to figure out the cause or a solution ... | | | ============================================================================ | | Mon, May 23, 2005 - 8:57:59 am | | glennmcc> | Which version of lsppp is Vladimir using. | | There was one particular version in-which the dialer _was_ 'broken'. | (I don't remember the v#) | | Anyways... the bug is now fixed and the dialer works perfectly. | | ---- my current ppp.log pasted-in via F3 --- | | LADSoft PPP Packet Driver Ver 1.0 Copyright (c) LADSoft | Dialing...Connecting to network...Authenticating...Negotiating IP linkage... | Connection open | Installed packet driver handler at vector 0x60 | PPPD status: IP address set to 4.252.90.232 | ____________________________________________________________________________ Š| | ============================================================================ | | Mon, May 23, 2005 - 9:06:00 am | | glennmcc> | --- pppdrc.cfg --- | 115200 | irq 4 | base 0xb800 | modem | crtscts | asyncmap 0 | namsrv | namsrv | user "glennmcc@cisnet.com" | passwd "*****" | ____________________ | | As you see... | no comport# (I have lsppp setup to use the base port address) | no DNS#s (lsppp gets them from the server) | | ---- lsppp.cfg --- | /M:atf50 | /x | /i:4 | /b:bx800 | /d:330-423-4005 | /B:115200 | /V:60 | /U:glennmcc@cisnet.com | /P:***** | _________________ | | --- my current environment --- | OS=OPENDOS | VER=7 | ARA=ON | PATH=\;C:\1BATCH;C:\1COMMON;C:\;C:\OPENDOS;C:\1ZIP;C:\QEMM;c:\borlandc\bin;c | OPENDOSCFG=C:\OPENDOS | BLASTER=A220 I5 D1 H7 T4 | MYIP=4.252.90.232 | REMIP=209.247.21.177 | NETMASK=0.0.0.0 | PEERMRU=1500 | DNS1=216.151.83.45 | DNS2=216.151.83.55 | COMSPEC=l:\command.com | TEMP=l:\temp | PROMPT=$d $b $t$_[OPENDOS 7.01] $P$G | ______________________________ | | ============================================================================ | Š| Mon, May 23, 2005 - 9:25:51 am | | ray> | Glenn, | | V says he's tried them all and even coresponded with the author | on the problem, so I don't think that's it. I like my new dialer | code, it essentialy formalizes what you are doing above, it just | expects you to put everything in lsppp.cfg or on the one 'connection' | line in arachne.cfg, but it doesn't work for Vladimir. No big deal, | the old code works fine for him, but I am very curious as to why | he needs to use miniterm as his dialer (which then requires all those | old keywords which could otherwise be removed). I'll show ya. | | | | ============================================================================ | | Mon, May 23, 2005 - 9:31:46 am | | ray> | | | // RAY: 04-04-12: Moved from 'adial.c' only used here. | // Create contents of $roura$.bat for dialing. | // RAY: 05-05-05: After .c33, start tinkering. | // No use of 'port', must specify base and irq ... life is tough. | // No use of cBREAK. | // No terminal. | // dospppd always true, removed. | // No need for hangup string. Should be a test to avoid calling dialer when | online | // Moved 'unlink' calls to end. | // Removed errorlevel " /?" and skip lines from buf. | // Removed 'terminal'. | // Rework buffer write into single sprintf. | // Forget pppdrc.cfg, seems not needed for lsppp. | | static char *near ArachneDial(void) | { | #ifdef EX1 | { | char buf[5 * IE_MAXLEN]; // Buffer for command sent to $roura$.bat | | ProcessForm(0, IE_NULL); // update Arachne.Cfg | | // NB 'dialer' is expendable, but 'connection' is used in guievent.c and | // atcpip.c for various things as well. | if (!sUI.connection) | { | puts("\nSORRY, YOU HAVE NO 'Connection' SET IN ARACHNE.CFG"); | delay(4000); | return (""); | } | sprintf(buf, "%Ns\n" Š| "@if exist IP-UP.BAT call IP-UP.BAT\n" | "@echo PPPD status: %Ns %MYIP%>>PPP.LOG\n" | ,sUI.connection, sUI.ip_grab); // "IP address set to" ... could be cMESSAGE. | | // There's gota be an easier way! | if ((!strcmpi(sUI.dialpage, p->htmlframe[p->activeframe].sCacheItem.URL) | || strstr(p->htmlframe[p->activeframe].sCacheItem.URL, "err_")) | && sArachne.scriptline == 0) | | strcat(buf, "@arachne.bat -o\n"); // Ends up at homepage, or shows error | else // see INIT.C c. L 630 and 735. | strcat(buf, "@arachne.bat -r\n"); // Reload previous page. | | // A TYPICAL 'buf' AS PRODUCED BY THE ORIGINAL CODE: | /* | @lh lsppp.exe /d:6046788350 /M:atf50 /n:2 /B:115200 /U:randrews ... | @if errorlevel 1 goto skip // What does this do? | @lsppp.exe /?>NUL // What does this do???? Help screen??? | :skip | @if exist IP-UP.BAT call IP-UP.BAT | @echo PPPD status: IP address set to %MYIP%>>PPP.LOG | @arachne.bat -r | */ | | // Buf2File("temp.txt", buf); | Outsf(cDIAL, cESC); | | // unlink("IP-UP.BAT"); // Used by lsppp.exe or epppd.exe to set env. variab | // unlink("PPP.LOG"); // Must delete to reset online timer but this | // is done in arachne.bat | return buf; | } | #else | { | ... old code ... | | ============================================================================ | | Mon, May 23, 2005 - 4:51:57 pm | | glennmcc> | And you called my code a kludge ???? | | ============================================================================ | | Mon, May 23, 2005 - 5:36:00 pm | | glennmcc> | Just back from looking-over you most recent post to AD (#1726) | | Hmmmm.... | | Many lines of code spread-out over several SRC files. | | versus Š| | 4 lines of code located in 1 SRC file. | | == no-brainer | | | ============================================================================ | | Mon, May 23, 2005 - 7:07:05 pm | | Joe> | Hi all, | | Glenn, if it makes you feel any better, and to put things in | perspective, I consider the whole LSPPPCFG package to be a kludge. | | | I'll have to look at Ray's code (when I have time!), before I can | give an opinion on it ... | | | ============================================================================ | | Mon, May 23, 2005 - 8:05:03 pm | | glennmcc> | Thanks for that Joe. | | It's nice to hear that someone else around here actually _likes_ a good klud | | | IMO, | Actually, when we really dig into her.... Arachne herself is the | culmination of hundreds of 'good kludges' combined together to make a | very useful and indispensible 'finished product'. | | ============================================================================ | | Mon, May 23, 2005 - 11:04:51 pm | | ray> | Glenn, | | I knew you'd react that way. Never mind, you fix does have the | virtue of beeing simpler to look at, but mine has the advantage | of being faster and actualy fixing the problem. Beware fixes that | create bottlenecks. | | | | ============================================================================ | | Tue, May 24, 2005 - 4:35:43 pm | | glennmcc> Š| You know what ? | | I am about sick-and-tired of this bickering that goes on between the two of | And I now intend to put a stop to it. | | I would like to ask a favor of you, if I may. | | I'll zip-up my current SRC code changes made between v1.83 and the | current one I'm working on. (v1.85) | | I'll make one last release in a few months using my version of the code | which is of-coarse the original SRC with the various changes made over | the last few years. | | Now for the favor.... If you could please 'meld' my 1.83-->1.85 changes | into your current itteration of the SRC and then zip-up the whole thing | to send back to me. | | My itteration of the SRC will then be zipped-up and set-aside. | | It's high-time I jumped onto your band wagon and started using the cleaned c | | | What you think ? ;-) | | ============================================================================ | | Tue, May 24, 2005 - 9:04:13 pm | | ray> | Glenn, | | Welcome aboard! But as I said some time ago, I've realized that | it is a good thing to keep both lines going as long as you've got the | patience to do it. But if you want to develop in the cleaned code, | that's great. As far as melding goes, that's basicaly stan | |y stando it again ! ×pfile:H:\down\adev0502.htm ------------------Z |$.----------------------------- ;* If command starts '@', it will be executed |and with maximum ; memory. ;* If command starts with [nn] string, it will be |without ; temporary leaving of Arachne if there is [nn] kB available for prog |command arguments you can use following macros. Instead of $1, $2, etc. ; lik |----------------------------------------------------------------------------- | ============================================================================ | | Fri, Apr 29, 2005 - 3:10:08 am | | Joe> | Hi all, | Š| Certainly, it would be relatively easy for LSPPPCFG to read the | appropriate stuff from arachne.cfg and use LSPPP for dialling. | However, I chose not to do this because 1) I wanted to retain the | terminal window capability (hi Ray;-) and 2) Because the LSPPP | dialler is less robust than Miniterm's (I have had complaints | about this from some users that tried to use LSPPP for everything). | | And, if I was convinced that nobody wanted the terminal window stuff | and that the LSPPP dialling was 100%, and changed LSPPPCFG to suit, | then most of the keywords above would still be required anyway. BTW, | presently, LSPPPCFG does use 'pppdrc.cfg', and whatever other similar | files are defined in the Arachne doc's (can't recall the details just | now, but as I recall there's an alternative file name and path search | options involved). Basically, LSPPPCFG uses the same files and settings | as EPPPD does, if that puts things in perspective. | | | ============================================================================ | | Fri, Apr 29, 2005 - 3:18:30 am | | Joe> | Hi Glenn, | | I forgot to mention ... I think you'll find that lsppp.exe doesn't | _require_ "bx800" syntax for its port address, merely it has poor | syntax checking and accepts this as "0xb800". You should find that | the correct syntax (ie. 0xb800) works just fine in both cases. | | | ============================================================================ | | Fri, Apr 29, 2005 - 3:24:32 am | | Joe> | Oh, about the WaitFor1, Response1, etc. keywords ... AFAIK, these | are used only for SLIP connections, right? (I certainly don't use | them with either EPPPD or LSPPP.) | | | ============================================================================ | | Fri, Apr 29, 2005 - 4:01:10 am | | gregy> | Question I have....since sometimes we (Wizzy, and maybe others) want | to dialup/connect/install packet driver...without loading Arachne (i.e. | I mean installing Arachne as "Online" instead of "dialup"....and we | know miniterm/epppd will do that, bacause Wiz has a package that does | that....since all of these things..... | ...can LSPPP be configured to do this, also.........???? | | | ============================================================================ | Š| Fri, Apr 29, 2005 - 8:02:10 am | | phantom> | Hi Gregy, | | Not sure what this "without loading Arachne" stuff is about, | particularly how dial-up is initiated, but the equivalent | would probably be something like : | | miniterm | lspppcfg | lsppp | | You'd probably do this via a batch file, because some ISP's only | give you a second or two between establishing a modem connection | and initiating PPP. You might also look at sub-pppd.bat for more | clues and ideas ... | | | | ============================================================================ | | Fri, Apr 29, 2005 - 9:32:18 am | | gregy> | Ah...I can get miniterm to dialup and connect, then do the PPP packet | install, then it drops....BUT....Wizzy has something that allows you | to miniterm-connect/PPP packet install/then fire up Arachne as if you | had an ethernet/NIC always-on connection, or fireup Lynx, or use ping, | traceroute, etc...... | ... | That's what I was asking about, if LSPPP was capable of dialing, | connecting, installing PPPpacketdriver, then sitting there waiting for | you to do something with some internet app.... | .... | hmmm...where do I find sub-pppd.bat....??? I don't find it in any | of my Arachne installs (up to 1.83).... | | | | ============================================================================ | | Fri, Apr 29, 2005 - 11:50:13 am | | phantom> | Hi gregy, | | Yep, I've used LSPPP as above, but not for quite a while. | Login to AD and check for the Lynx4AU packages, mentioned | on the main page. The let you switch between Arachne and | Lynx. | | | ============================================================================ | Š| Fri, Apr 29, 2005 - 12:01:55 pm | | phantom> | In other words, you can use Arachne as your main TCP/IP client, | including its dialling capabilities, exit whenever you like to | run other clients, such as Lynx, then return to Arachne to | continue or to finish your online session. | | | ============================================================================ | | Fri, Apr 29, 2005 - 12:14:22 pm | | phantom> | In case you hadn't guessed, I prefer to do my dialling/connection | via Arachne, only resorting to more manual methods for diagnostic | purposes. | | Oh, from memory the Lynx4AU packages are in the Tools directory | in the AD Files area. If not, check the Links area for my web | site, also for the Lynx site. This stuff will give you an idea of | just how flexible this TCP/IP connectivity can be, using either | EPPPD or LSPPP. | | | | | ============================================================================ | | Fri, Apr 29, 2005 - 12:24:40 pm | | phantom> | Final hint for the night ... If you've played with Miniterm and | LSPPP but get drop-out, I'd suggest starting from scratch with | a standard EPPPD configuration. Once that's working, it takes | just a few seconds with the LSPPPCFG package to convert to a | working LSPPP configuration. Then you'll be able to use Miniterm | and LSPPPCFG/LSPPP manually too, but you must first get it to | work via Arachne. BTW, the sub-pppd.bat file is generated when | you install the LSPPPCFG package. | | | ============================================================================ | | Fri, Apr 29, 2005 - 2:07:48 pm | | glennmcc> | No problem dialing/establishing PPP without Arachne. | | In arachne.cfg | | Connection Ready | | AT the DOS prompt while in the arachne_main dir | Š| lh lsppp.exe>ppp.log | | If lsppp.exe and lsppp.cfg are not in the arachne_main dir....... | | cd \lsppp | | lh lsppp.exe>\arachne_main\ppp.log | | There you go............. | | You are connected and ready to go when you decide to run arachne.bat | | ============================================================================ | | Fri, Apr 29, 2005 - 2:10:07 pm | | glennmcc> | Joe, | | Thanks for the 'heads up' | | I just tried it and you are right... 0xb800 works perfectly in lsppp.cfg | | | ============================================================================ | | Fri, Apr 29, 2005 - 5:51:23 pm | | ray> | Lots to think about. From what I hear, it's back to lspppcfg for me | since it maintains compatability with existing keywords, although I do | like the idea of an alternate lspppcfg that was built to include | dialer services as well for folks like me who don't have any | dialer issues with lsppp and would like to just dump miniterm. No | biggy though, as she is now, all is well. | | | | ============================================================================ | | Fri, Apr 29, 2005 - 5:54:41 pm | | glennmcc> | IIRC, | There was one particular release of lsppp.exe in-which the dialer | functions were 'broken'. | | As far as I know... everything is fine in all other releases. | | | ============================================================================ | | Fri, Apr 29, 2005 - 6:29:02 pm | | ray> Š| Hmmm, then maybe there is nothing wrong with the current versions and | we can persuade Joe to update lspppcfg. If I followed the previous | remarks, the only remaining difficulty would be that one couldn't use | 'terminal' mode. That's never done anything for me although others | might have found it useful. | | | ============================================================================ | | Fri, Apr 29, 2005 - 7:36:47 pm | | glennmcc> | terminal mode is only good for 'diagnostics' | | If you are having trouble logging on, you can follow the process | step-by-step to where it goes wrong. | | That is exactly the reason why I put this on pppglenn.htm which has now | been renamed as pppenhan.htm | | _ show terminal (diagnostic mode) | | ============================================================================ | | Fri, Apr 29, 2005 - 8:13:40 pm | | ray> | Ya, I know what it's for it's just a question of how useful it | still is. | | ============================================================================ | | Fri, Apr 29, 2005 - 8:48:51 pm | | glennmcc> | It's _very_ useful every time there is a problem with login. | | | ============================================================================ | | Fri, Apr 29, 2005 - 8:50:46 pm | | ray> | Well, then we still need the old keywords, no? | | | ============================================================================ | | Fri, Apr 29, 2005 - 9:02:15 pm | | glennmcc> | Some yes, some no. | | If you'll notice the list of keywords I felt could be removed....... | niether 'UseTerminal' nor 'TerminalWindow' is among them. Š| | | Port | Irq | Base | Mode | Speed | InitString | DialString | PhoneNumber | WaitFor1 | Response1 | WaitFor2 | Response2 | WaitFor3 | Response3 | PPPusername | PPPpassword | | [tcp/ip] | NameServer | AltNameServer | | ============================================================================ | | Fri, Apr 29, 2005 - 9:17:25 pm | | glennmcc> | We don't need to use miniterm.exe as our terminal. | | I am logged-on right now using lsppp.exe as my 'terminal' | with these settings in arachne.cfg | | Connection @lh lsppp.exe /?>NUL | Dialer lsppp.exe /?>NUL | TerminalWindow @lh lsppp.exe | UseTerminal Yes | Autologin No | IP_Address BOOTP | | | ============================================================================ | | Fri, Apr 29, 2005 - 9:32:58 pm | | glennmcc> | We will of-coarse need to use either miniterm or some other 'real' terminal | program to do a 'manual login' in which we enter our username and password | via the keyboard at the needed points during the login process. | | ============================================================================ | | Sat, Apr 30, 2005 - 7:08:57 am | | gregy> Š| Thanks, guys...for the comprehensive discussion... | .. | I think you've give me enough hints and tips to get me through the | trauma.... | | | | ============================================================================ | | Sat, Apr 30, 2005 - 7:14:08 am | | gregy> | One of the things I wanted to do, was to just get an independent | LSPPP dialup/connect install working so I can check out Glenn's | report that his ISP actually gives him his DNSes (server-supplied | DNS) when he connects, just as we always knew our ISPs (most of them, | anyway) give us dynamically-assigned personal IPnumbers to use during | our online connection session. | | Miniterm/EPPPD cannot do that (or don't, anyway) and Glenn gave a | listing of his environment variables showing that LSPPP stashed two | DNSx numbers in there. If my ISP is doing this, also, that would | probably explain why the Shiva Dialer/Winsock could actually still | access DNSes even though I had reset its two DNSvariables to 0.0.0.0 | | | | ============================================================================ | | Sat, Apr 30, 2005 - 7:52:21 am | | ray> | Glenn, | | But wouldn't the PhoneNumber still be needed to use miniterm? | Hell, it's the dialer no? | | | ============================================================================ | | Sat, Apr 30, 2005 - 2:36:39 pm | | glennmcc> | Yes, it would. | | But what I was talking about was elimiating miniterm. | | | | ============================================================================ | | Sat, Apr 30, 2005 - 2:40:13 pm | | glennmcc> | All of these are only needed when using miniterm. Š| | Port | Irq | Base | Mode | Speed | InitString | DialString | PhoneNumber | WaitFor1 | Response1 | WaitFor2 | Response2 | WaitFor3 | Response3 | | | These are only needed when using epppd | | PPPusername | PPPpassword | NameServer | AltNameServer | | | When we use lsppp as both dialer and packet driver we no longer | need any of those lines in arachne.cfg | | All of that information is now in lsppp.cfg instead. | | | | | ============================================================================ | | Sat, Apr 30, 2005 - 6:31:20 pm | | ray> | Yeah, I just meant if Miniterm was still needed for this terminal stuff. | | | ============================================================================ | | Sat, Apr 30, 2005 - 6:48:18 pm | | glennmcc> | It's not needed for 'terminal stuff'. | | As I showed... lsppp.exe can be used as the 'TerminalWindow' | | Miniterm (or an other terminal program we choose), | can be used for 'manual login'. | | But since almost all ISPs have now disabled the option of manual login | it is now a moot point. Š| | Virtually all ISPs now _require_ PAP/CHAP login and manual login will not wo | | ============================================================================ | | Sat, Apr 30, 2005 - 6:52:56 pm | | glennmcc> | The only thing we need the terminal window for anymore is so that we can | see what's going on during dialup, connect, PAP/CHAP login, PPP negotiation. | | We only need to see this stuff when we are having a problem logging in. | | By choosing to use the terminal window, | we can diagnose what the problem is. | | That's why I marked it (diagnostics mode) on pppenhan.htm | | | ============================================================================ | | Sat, Apr 30, 2005 - 8:48:30 pm | | ray> | Ok, got it ... more or less ;-) ... but if actual manual login | is obsolete, then the terminal window is just a viewer, no? | And if that's true, can't its output be logged to a file for viewing? | If so, then old miniterm would truely be history, no? | | BTW a bit of a milestone here: core.exe 397,856 | | | ============================================================================ | | Sat, Apr 30, 2005 - 8:51:59 pm | | ray> | Oh, almost forgot, you don't suppose that problem of Bastiann's is | due to your cutting back on static modules do you? If I was you, I'd | send him a core with the modules as before, and see if that fixes his | problem since that could prove to be an important bit of information | as to what can and can't be overlayed. And we both want to fart | around with that. | | | | ============================================================================ | | Sat, Apr 30, 2005 - 9:55:10 pm | | glennmcc> | It already _is_ logged to a file. | | That's the reason for lsppp.exe>ppp.log | Š| | | | ============================================================================ | | Sat, Apr 30, 2005 - 10:05:25 pm | | ray> | So then miniterm really is rather marginal, no? | | | ============================================================================ | | Sun, May 01, 2005 - 2:01:40 pm | | glennmcc> | yes.... 'marginal'. | | Maybe we still need it..... maybe not. | | ============================================================================ | | Sun, May 01, 2005 - 8:07:25 pm | | glennmcc> | I've been doing some experimenting this evening and I just did a | successful fresh compile of v1.66b ......... | | | are you ready for this ? | | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - Š| - | - | - | - | | for Linux svgalib :) | | | ============================================================================ | | Sun, May 01, 2005 - 9:42:14 pm | | ray> | He he, did you really get all the graphics calls migrated? Sounds too | good to be true, but if it is as you say that's f***ing marvelous, | pardon my Canadian French. We now have a doorway to the future. | | | ============================================================================ | | Mon, May 02, 2005 - 12:11:01 am | | Joe> | Glenn, | | I think Port, Irq, Base, Speed may still be required without Miniterm. | (Perhaps Speed is optional. IIRC, one either uses Port or Base/Irq, | not both.) | | ============================================================================ | | Mon, May 02, 2005 - 12:17:16 am | | Joe> | Gregy, | | Dynamic DNS is standard fare these days and has been for quite a | while. It's been supported by LSPPP for quite a while too, and | IIRC, has always been the default configuration with LSPPPCFG | (since ca. 2000). In fact, a number of ISP's in Oz will not allow | you to select a different DNS, only whatever their server tells | your TCP/IP application. The way it works is that 0.0.0.0 is not | a valid IP, so the server says to the application, "No, use these | IP addresses for DNS instead". The ISP's I use here will say that | for non-zero IP addresses too. | | | | ============================================================================ | | Mon, May 02, 2005 - 9:36:55 am | | glennmcc> | Hi Joe, | Š| Nope.... not needed in arachne.cfg _if_ miniterm/epppd are not used. | | When we use lsppp as both dialer and packet driver, all of the needed | information is in lsppp.cfg instead of arachne.cfg | | ---- my current settings in arachne.cfg --- | | [profile] | Profile 4234005L.ACF | Connection @lsppp.exe /?>NUL | Hangup @termin.com 0x60>NUL | | [dialer] | DialPage file:pppenhan.htm | | Dialer @lh lsppp.exe>ppp.log | UseTerminal No | Autologin No | | [tcp/ip] | IP_Address BOOTP | IP_Grab IP address set to | Gateway 0.0.0.0 | AltGateway 0.0.0.0 | Netmask 0.0.0.0 | TCPconfig PPPTCP.CFG | ___________________________________________ | | | --- my current settings in lsppp.cfg --- | /M:atf50 | /x | /i:4 | /b:bx800 | /d:330-423-4005 | /B:115200 | /V:60 | /U:glennmcc@cisnet.com | /P:***** | ________________________________________ | | As you see... | all of the lines I mentioned have been deleted from arachne.cfg | and the ones that are needed for lsppp are over in lsppp.cfg instead. | | ============================================================================ | | Mon, May 02, 2005 - 9:46:40 am | | glennmcc> | When using miniterm to dial _and_ a special base port address is needed. | (such as in my case of 0xb800) | | _both_ port # _and_ base port address must be used. | Š| This is due to the fact that 'port 0' is the signal to miniterm.exe to | use 'base #' instead. | | If we delete the port line (or comment it out with ; preceeding it), | miniterm will not dial because it uses 'port 2' by default when | 'port #' is missing from arachne.cfg | | ============================================================================ | | Mon, May 02, 2005 - 8:15:12 pm | | ray> | I done a good thing today: I've created the keyword "Experimental" | which let's the user turn off anything experimental with one keyword. | This permits me have a 'stable' core and an 'experimental' core | at the same time. I'd give it some thought Glenn, it would let | your customers back out of anything iffie and thus allow you to | experiment more without having to worry about new bugs poping up | and having to scramble to patch it. | | | | ============================================================================ | | Mon, May 02, 2005 - 9:11:52 pm | | glennmcc> | I prefer this method..... | | #ifndef GLENNMCC | //experimental stuff | #endif | | | This way the experimental stuff does not get compiled into the release | version but only into the ones I compile for myself using glennmcc.prj | | | | ============================================================================ | | Mon, May 02, 2005 - 9:13:19 pm | | glennmcc> | Oops, typo... | | That actually reads "#ifdef GLENNMCC" not "#ifndef GLENNMCC" | | ============================================================================ | | Mon, May 02, 2005 - 9:17:16 pm | | glennmcc> | Since I always use nokey.prj for the release versions. | This method also works great. Š| | #ifndef NOKEY | //experimental stuff | #endif | | | ============================================================================ | | Mon, May 02, 2005 - 9:53:48 pm | | ray> | Yeah, that's cool too, but it serves a slightly different need. | I'd say my method is more for code that is considered ready for | prime time, but the config option would be used as a failsafe in case | of some unforseen problem, or just in case some new feature wasn't | liked. I'd use the conditional for very raw experimental stuff | or for some custom build or other. | | BTW any word from Bastiaan on that stutter? I can't remember if I | mentioned it or not, but I'm running right now with an overlay | buffer forced to 128K, and the improvement in speed is quite nice. | It will be interesting to compare that with various modifications | to the number of modules made static. | | | ============================================================================ | | Tue, May 03, 2005 - 6:55:52 am | | phantom> | Hi Glenn, | | Yes, you are correct about Port being required in any case, if only | to indicate "use Base and Irq instead", ie. the special value 0. I | haven't tried simply omitting it, perhaps this will have the same | effect? | | Now as for all those keywords being "redundant" without Miniterm, | the problem here is a difference in perspective. I consider it | Arachne's job to control its dialler and packet driver, in which | case this is either done directly (for Miniterm/EPPPD) or indirectly | (for Miniterm/LSPPPCFG/LSPPP). If we add lsppp.cfg support to Arachne, | then once again, these keywords will be used directly to control | LSPPP. I prefer all the configuration screens and ACF's and wizards | to work in configuring these settings, not a separate configuration | file (ie lsppp.cfg) that the user must manually edit if they want to | use LSPPP. One of my primary goals in writing LSPPPCFG was to eliminate | the need to edit any file whatsoever, to use LSPPP with Arachne in | exactly the same manner as for EPPPD. | | | ============================================================================ | | Tue, May 03, 2005 - 5:22:29 pm | Š| glennmcc> | re: | "haven't tried simply omitting it, perhaps this will have the same effect?" | | Nope.... omitting it causes miniterm to default to 'port 2' | | As for the rest......... You are 100% correct. | | If you can modify lspppcfg.exe to 'grab' all of those lines from | arachne.cfg and place them into lsppp.cfg in thier correct formats. | | We will then have 'perfection'. ;-) | | | | | ============================================================================ | | Wed, May 04, 2005 - 12:41:24 am | | Joe> | Hi Glenn, | | Port 2, eh? | | Yuck!!! | | ============================================================================ | | Wed, May 04, 2005 - 3:48:54 pm | | glennmcc> | Yep..... Yuck is right. | | This is just one more reason to DITCH miniterm and go with lsppp for | _both_ dialer and packet driver. | | ============================================================================ | | Wed, May 04, 2005 - 5:09:23 pm | | ray> | Yup, let's stick with one proven product and make it work. | | | | ============================================================================ | | Wed, May 04, 2005 - 5:56:20 pm | | ray> | What is 'chk-optn' as mentioned in 'sub-pppd.bat'? I can't find | that as a command or even as a string anywhere. | | Š| ============================================================================ | | Thu, May 05, 2005 - 6:00:26 pm | | Joe> | Ray, | | 'chk-optn' is a utility from the LSPPPCFG package which is used | to check "option settings" in '$roura$.bat'. It is used to determine | what command line option the LSPPPCFG package should use to invoke | 'arachne.bat'. This feature was required because your Arachne | configuration fetched e-mail first, rather that going to a default | (home) web page, early versions of the LSPPPCFG package assumed the | latter and always used the "-o" command line option. | | | ============================================================================ | | Thu, May 05, 2005 - 8:50:50 pm | | ray> | Thanks Joe, | | I found it using a wider search, I had managed to loose it in my | working copy of Arachne, so I was running without if for some time | without any noticable problems, but as you say, this seems to | be a responce to a specific situation. | | | | ============================================================================ | | Sat, May 07, 2005 - 8:25:38 am | | ray> | What's the option that makes the arrow keys move the cursor instead | of scrolling the document? | | ============================================================================ | | Sat, May 07, 2005 - 10:54:03 am | | ray> | Joe, Glenn, | | I'm having fun butchering the dialer code, got it cut down to | about 1/3 the original size. There will be mistakes in this, but | it's been instructive. A question: what are the errorlevel values | returned by lsppp.exe, what, when and why? | | | ============================================================================ | | Sat, May 07, 2005 - 2:45:31 pm | Š| glennmcc> | Realtime smooth scrolling ("Netscape") | | | ============================================================================ | | Sat, May 07, 2005 - 8:25:03 pm | | Joe> | Me confused! | | I thought the answer was : the shift key! | | ============================================================================ | | Sat, May 07, 2005 - 8:28:54 pm | | Joe> | Ray, | | Well, early versions of LSPPP did not set the errorlevel | on exit - it was random (in the sense of being meaningless). | | On my suggestion (insistance;-), DL fixed this, however I'm | not sure without studying the code if there's specific levels | used, or if it's just a matter or zero vs. non-zero. | | ============================================================================ | | Sat, May 07, 2005 - 8:31:47 pm | | Joe> | Typo ... that should have read as : | ~~~ a matter of zero vs. non-zero. | | | ============================================================================ | | Sat, May 07, 2005 - 10:55:11 pm | | glennmcc> | Hi Joe, | | You not confused..... shift is method #2 ;-) | | Turning off Smooth scrolling eliminates the need for holding the shift | key to get the arrow keys to move the mouse pointer. | | ============================================================================ | | Sun, May 08, 2005 - 6:10:48 pm | | ray> | Glenn, Tx, I always forget that one. BTW, why doesn't some clever | hotshot combine the two so that if you are using the cursor movement Š| setting, once the cursor hits the bottom of the screen the page starts | scrolling? This would be a third option for 'Smoothscroll' I figure. | | Joe, | | The reason I ask is that I'm trying to figure out this, which is | the contents of $roura$.bat when the dialer is invoked: | | | /* | @lh lsppp.exe /d:6046788350 ..... | @if errorlevel 1 goto skip | @lsppp.exe /?>NUL | :skip | @if exist IP-UP.BAT call IP-UP.BAT | @echo PPPD status: IP address set to %MYIP%>>PPP.LOG | @arachne.bat -r | */ | | The way it looks, an errorlevel of one is OK and it skips the | next line which as far as I can tell does nothing useful anyway. | I've butchered all of that out anyway to see what happens and it seems | fine. Next I'll try to incorporate what I think your lspppcfg does | inside the dialer and see what I can come up with. | | | ============================================================================ | | Sun, May 08, 2005 - 6:20:49 pm | | ray> | Oh, Glenn, I noticed that rotten old smartdrv solved B's problem | but I'd like to know which file is the culprit, ie the file that | needs to be reloaded so often, that would be a good candidate for | static link no? I'm all the more curious since when you sent him | a build with the original plan, it didn't help. Perhaps there is some | file that never was static that should be. | | | | ============================================================================ | | Sun, May 08, 2005 - 6:47:41 pm | | glennmcc> | It's in the Internal Editor stuff.... | | The IE*.c files. | | BRB | | | ============================================================================ | | Sun, May 08, 2005 - 7:54:14 pm Š| | ray> | Hmmm, the obvious suspects. | | ============================================================================ | | Sun, May 08, 2005 - 8:19:54 pm | | glennmcc> | Nope, I was wrong.... that was not it. | | But then I can't duplicate Bastiaan's "stuttering" problem using any of | the release versions of core.exe | | | ============================================================================ | | Sun, May 08, 2005 - 9:23:57 pm | | ray> | Could be some one-off anomaly not worth chasing or maybe the | speed of his machine is a factor -- IIRC it was rather slower than yours | but then there are several people running relatively slow machines | with no trouble. | | ============================================================================ | | Mon, May 09, 2005 - 10:46:14 am | | gregy> | Joe....thanks for the confirmation. My ISP swears they don't | auto-supply the DNSnumbers...but...the operation of that Win3.1 Shiva | Dialer says they do......that is why I want to investigate this using | LSPPP. | ... | However....please do not refer to this as "Dynamic DNS"...if you google | that, you will find that it refers to something very different that | what we are talking about. | ... | Google "Server supplied(provided) DNS" (as Glenn did) and you will | come up with what we are talking about. This "INET Terminology" problem | has bit us a couple times in the past (that I can remember)....people | talking about one thing, while using terminology that refers to quite | another thing. | | That's what thru me off, and perhaps thru my ISP's tech people off (I | havent' gone back to them with the term "server supplied DNS"..so maybe | they're not so clueless). But, those two different terms are what | threw me off, and got me going in the wrong direction. | ... | he he...thank GOD we gots lotsa "redirectors" around here... | | | ============================================================================ | Š| Mon, May 09, 2005 - 10:47:44 am | | gregy> | Glenn, Ray.....I could not reproduce Bastiaan's stutter using 1.83 on | this Cyrix 486dlc 40MHz box, so I don't think it's due to chip speed. | .. | | | | ============================================================================ | | Mon, May 09, 2005 - 2:50:04 pm | | ray> | Gregy, as one last test, try running it with your cache disabled, if | that doesn't reproduce the problem then it would seem to be some | local anomaly of Bastiann's | | | ============================================================================ | | Mon, May 09, 2005 - 4:59:33 pm | | glennmcc> | I am on line right now with v1.83 core.exe running 100% on diskette. | No diskcache loaded at-all. | stutter...stutter... stutter... | BRB with the compile I made for bastiaan that has all of modules back | the way they originally were. (static vs overlayed) | | ============================================================================ | | Mon, May 09, 2005 - 5:11:48 pm | | glennmcc> | Now I'm back with the core.exe I compiled on May 1st | Still stuttering :( | BRB | | | | ============================================================================ | | Mon, May 09, 2005 - 5:21:27 pm | | glennmcc> | Now let's see ............ | | Here's a string a of random characters typed very quickly to see what happen | | .kjndvc:SJHF;OIENGF/AKFLJHT90[2QU536QPAKINBVZ[9YUHq]nBslg'dap]t90-qjh32yt' | ;oasmnbvaxjlc]opjtgA;MLGZx:MLBVASKTRP.,mvn:LKXNVCKLNfsKLANFDLKanf/;KLznFk | lansfd/;KLANF/;klansfkl;ANSF/;KLandflkNADFklandklfANSDFKL;anfl;kasndgfkla | nsklgnasklgfdnadsklgna/;lskbvnasklngd | Š| We got it...... NO STUTTERING AT-ALL !!! :))) | | ============================================================================ | | Mon, May 09, 2005 - 5:24:40 pm | | glennmcc> | Comparing my files I find that I screwed-up back on May 1st and still | had the .PRJ in-place which had some additional modules overlayed. | | This compile I just made now has everything back the way it was. | | ============================================================================ | | Mon, May 09, 2005 - 5:27:59 pm | | glennmcc> | I will upload a ZIP to my site in a few minutes with everything that's | on this diskette. | | This sucker is a perfect way of testing disk activity. | | | ============================================================================ | | Mon, May 09, 2005 - 5:45:13 pm | URL | | glennmcc> | Unzip to A:\ with........ | | pkunzip -d diskette.zip a:\ | | Also place your own ARACHNE.CFG, ARACHNE.PCK and LSPPP.CFG | over onto the diskette. | | Also copy MINITERM.EXE and EPPPD.EXE onto the diskette | if you use them for dialer/packet driver instead of LSPPP.EXE | | Alter the path lines in ARACHNE.CFG to point to A: as-well. | (make new dirs as needed on A:\) | | ============================================================================ | | Mon, May 09, 2005 - 6:19:42 pm | | ray> | Cool, all this is grist for the mill as far as understanding goes. | Please advise when you find the exact culprit. | | | ============================================================================ | | Mon, May 09, 2005 - 9:32:32 pm | Š| gregy> | ummm...Ray...I have to report a "local anomalie" of my own.... | ... | I doan hav no steengkin diskcache........In this DRDOS 6.0 it | is PCKWIK and I have all reference to it removed from my boot files. | ... | I've had it this way ever since I figured out that the ONLY way to | run Arachne is from a RAMDisk.... | .... | | | | | ============================================================================ | | Tue, May 10, 2005 - 3:14:03 am | | Joe> | Glenn, | | Thanks for the info about cursor keys and Smooth Scrolling, | I had *no* idea they were in any way related! | | ============================================================================ | | Tue, May 10, 2005 - 3:16:24 am | | Joe> | Ray, | | Lots of fun with your experimentation. FYI, I don't think LSPPPCFG | in its present version is quite what you need. I have been giving | some thought to how I should enhance it for LSPPP dialling, I'll | look into that after doing the "clean code" version of 1.83J1 ... | | ============================================================================ | | Tue, May 10, 2005 - 3:27:39 am | | Joe> | Gregy, | | Thanks for the "heads up" on the "Dynamic DNS" jargon. It seems they | (who?) want to (re?)define this as "DNS support for Dynamic IP", so | no wonder there's some confusion here. Silly people, huh? | | | ============================================================================ | | Tue, May 10, 2005 - 1:31:29 pm | | ray> | Gregy, | | Whaaa, are you a wise-guy or sometheeen? ;-) Š| | I forgot you're a ramdisker. I suppose that would have more or less | the same effect on speed issues as would cacheing, or better, so | there's little to be learned there. | | | ============================================================================ | | Tue, May 10, 2005 - 2:04:13 pm | | ray> | Joe, | | Roger that. FWIW, I'll post my boiled down dialer code on AD. | | | ============================================================================ | | Tue, May 10, 2005 - 6:28:18 pm | | ray> | What's our established minimum free memory? I've been running for | some time with my core set to be rather glutonous so as to leave only | about 106K free, and it seemed ok until I ran a script and crashed. | | | | ============================================================================ | | Tue, May 10, 2005 - 7:28:13 pm | | glennmcc> | IMO, | | 450 is the absolute min (_IF_ one is able to load the packet driver high) | 500 to 550 when conv mem must be used for the packet driver. | | ============================================================================ | | Tue, May 10, 2005 - 7:32:37 pm | | glennmcc> | My current stats (lsppp loaded high) | | DOS mem (bytes) 200672 | | --- with core.exe 'dumped' (memc.dgi) --- | | DOS 91,280 ( 89K ) 16490 | QEMM386$ 768 ( 0K ) 300 | COMMAND 1,936 ( 1K ) 790 | FREE 642,928 ( 627K ) 9CF70 | Largest executable program: 642,912 ( 628K ) | Total Free DOS memory: 720,864 ( 704K ) | Š| ---- with core.exe still in memory (memcore.dgi) --- | | DOS 91,280 ( 89K ) 16490 | QEMM386$ 768 ( 0K ) 300 | COMMAND 1,936 ( 1K ) 790 | CORE 442,768 ( 432K ) 6C190 | command 7,200 ( 7K ) 1C20 | FREE 192,960 ( 188K ) 2F1C0 | Largest executable program: 192,944 ( 188K ) | Total Free DOS memory: 270,896 ( 265K ) | | | ============================================================================ | | Tue, May 10, 2005 - 7:37:02 pm | | glennmcc> | by "ran a script"... do you mean an .ASF ? | | If so... | I have found that .ASFs containing 'reload:' usually require no less than | DOS mem 150000 to prevent a crash. | | IMO, | .ASF handling needs a 'major overhaul'. | | | ============================================================================ | | Tue, May 10, 2005 - 9:22:35 pm | | ray> | Yes to all. I think the magic number might be 128K free (core.exe loaded | of course) I did some crude memory destroying tests and found that | both our cores are ok with 'reloading' ASF's if 128K+ is available. | That's the size of two 64K malloc() buffers so it seems likely. | So what is it about ASF's that require such huge buffers? If I can | find a more accurate way of filling up my memory, I'll run more | exact tests. | | | ============================================================================ | | Thu, May 12, 2005 - 9:22:33 pm | | ray> | Fixed a wee bug that's been bothering me for some time: | You know how in a 'compose mail' screen, if you move the mouse | directly down from the input area onto a button if forgets it's color? | I think I've bagged it. | | And, in the process I've added a third mouse color for when it's on | a text area. | | Š| ============================================================================ | | Thu, May 12, 2005 - 10:57:55 pm | | ray> | So Glenn, what about those colours? What's the pattern there? | | | | ============================================================================ | | Fri, May 13, 2005 - 5:01:30 pm | | glennmcc> | ????? | | I don't understand your question. | | ============================================================================ | | Fri, May 13, 2005 - 5:56:53 pm | | ray> | If you set your mouse colors to '2 2' your cursor will turn RED. | If you use '3 3' you get GREEN ... this is hard to explain using | either IRGB or RGB or RRGGBB or anything else that I can think of. | It doesn't matter a whole lot of course, one can simply take it on | faith that 3 == GREEN, but I'd like to know why. | | ============================================================================ | | Fri, May 13, 2005 - 8:19:26 pm | | glennmcc> | Here's what I've just found by checking 0 to 15 in MouseColors | | 0 black | 1 dark blue | 2 dark red | 3 dark green | 4 dark purple | 5 dark aqua | 6 dark yello | 7 light gray | 8 dark gray | 9 bright blue | 10 bright red | 11 bright green | 12 bright purple | 13 bright aqua | 14 bright yellow | 15 white | | | Checking into config.c I find that the CHAR value of MouseColors is Š| the int of MouseColors after being converted into a longint | | (int) strtol(value, n10) | | MouseColors 0 3 == the int of the longint 03 in base 10 | | I think that there MUST be a better way. | | We are getting 16 color possibilities..... but theye don't match | the 'standard 16 color table' | | 0 BLACK | 1 BLUE | 2 GREEN | 3 CYAN | 4 RED | 5 MAGENTA | 6 BROWN | 7 LIGHTGRAY | 8 DARKGRAY | 9 LIGHTBLUE | 10 LIGHTGREEN | 11 LIGHTCYAN | 12 LIGHTRED | 13 LIGHTMAGENTA | 14 YELLOW | 15 WHITE | | | | ============================================================================ | | Fri, May 13, 2005 - 8:21:04 pm | | glennmcc> | Gregy !!!!!!!! | | Fix this, will ya please ? | | (int) strtol(value, & newvalue, 10) | | ============================================================================ | | Fri, May 13, 2005 - 9:19:50 pm | | ray> | Yes, that's my point, the lopif functions have their own strange | ideas about colors. I used to be just vaguely disturbed by it | and now I know why, since one would expect the standard values to | be used. BTW, the strtol isn't the problem since I long ago | started using a simpler parsing routine for those values, you can | plug numbers directly into the functions and it makes no difference. | | | Š| ============================================================================ | | Fri, May 13, 2005 - 9:21:39 pm | | ray> | What puzzles me even more is that exactly half of the results we | see *do* match the standard table -- '14' is bright yellow just | as one would expect it to be. Puzzling, no? | | | | ============================================================================ | | Fri, May 13, 2005 - 10:45:16 pm | | glennmcc> | Ah HA !!! | | So, the broblem is not in the SRC file(s) of core.exe | | It's somewhere in the SRC files of HIXLOV.LIB | | ============================================================================ | | Sat, May 14, 2005 - 8:31:05 am | | ray> | Exactly, this ain't our problem, it's Haro's problem. Without Michal, | it's hard to go further except to just be aware of the discrepancy. | | | ============================================================================ | | Sat, May 14, 2005 - 2:53:31 pm | | glennmcc> | OK, sounds good to me. | | Suggestion.... | | How about if we include that color table in the documentation ? | | BTW, | It applies to HTML BGcolor and text color as-well. | | Arachne.cfg variable | | Colors n n | | Works in exaclty the same manner as does | | MouseColors n n | | Therefore this color table applies to both of those CFGvariables | Š| 0 black | 1 dark blue | 2 dark red | 3 dark green | 4 dark purple | 5 dark aqua | 6 dark yello | 7 light gray | 8 dark gray | 9 bright blue | 10 bright red | 11 bright green | 12 bright purple | 13 bright aqua | 14 bright yellow | 15 white | | | | ============================================================================ | | Sat, May 14, 2005 - 3:02:26 pm | | glennmcc> | Oops... my bad. :( | | Colors n n | | Does not refer to HTML BGcolor and text color | | It refers to the text color and BGcolor of text input fields | | Colors 0 15 | | Makes this text area, the input fields for Nick:, Email:, and URL: | and the URL bar showing the current page URL .... black text on white BG | | ============================================================================ | | Sat, May 14, 2005 - 7:00:27 pm | | ray> | Yup some documentation would be apropriate, we both have the facts | now. BTW, the affected functions AFAICT are: x_setcolor(), x_setfill() | and x_yncurs() (names from memory!) there may be others but I haven't | found them. I now have defines in place for all calls (which is what | I set out to do several day ago ;-) | | BTW I have a bug report: Go to the move mail page, pull down the | menu for other folders and move the mouse rapidly off the menu and | back on to neutral territory, what happens? | | | ============================================================================ | Š| Sat, May 14, 2005 - 9:12:41 pm | | ray> | You got no problems there? | | | | ============================================================================ | | Sat, May 14, 2005 - 11:20:49 pm | | gregy> | Glenn, I will try to put the new-style board up this next week... | ... | I am sorry that I have been so sorry..........this winter I | have not been able to do or complete very many things... | | | ============================================================================ | | Sun, May 15, 2005 - 8:39:39 am | | ray> | Ha, well spring planting can't be far off and you won't have any time | then ;-) | | | ============================================================================ | | Sun, May 15, 2005 - 2:49:11 pm | | glennmcc> | Thanks gregy. | | BTW, | | Please be sure to 'archive' this board before making the changes. | | | ============================================================================ | | Mon, May 16, 2005 - 10:34:35 pm | | gregy> | yes...I'll d/l the whole thing, convert to print, then split it up | by quarter-years, or something... | .. | | | ============================================================================ | | Tue, May 17, 2005 - 9:45:53 pm | | ray> | Glenn, Š| | You'll have to kill me. Downloaded 1.83, poached your fix for that | SELECT thing and she works great. Inspired by your insights, I | went a little further and managed to simplify the code quite a bit. | Still, it's just a hack; why should the SELECT tag cause an endless | loop? OnMouse() should be able to handle that, or, second best, | the specific call to MouseOn() that's causing the fault should have | the filter aplied there. | | | ============================================================================ | | Wed, May 18, 2005 - 12:10:21 pm | | glennmcc> | Hey, hey, hey, watch it there bud !!! | | That's _my_ hack you're jabbing at. | | Yes, it's a hack... and I'm damned proud to be a hacker. | | | ============================================================================ | | Wed, May 18, 2005 - 1:03:23 pm | | ray> | Sure, it's a hell of a lot better than mine was but still a hack. | All I'm saying is we should keep on the lookout for a fix that's | closer to the heart of the problem. | | | ============================================================================ | | Wed, May 18, 2005 - 4:40:22 pm | | glennmcc> | Now he wants to stab me in the heart. | | Sheesh...... sometimes a guy just can't win. | | | | ============================================================================ | | Wed, May 18, 2005 - 4:43:47 pm | | glennmcc> | But seriously now... | | If it ain't broke don't fix it. | | So if the fix ain't broke don't fix it. | | Š| ============================================================================ | | Wed, May 18, 2005 - 4:45:42 pm | | glennmcc> | If you feel like fixing something that REALLY needs fixing..... IgnoreJS | | | | ============================================================================ | | Wed, May 18, 2005 - 5:47:20 pm | | Joe> | Hi Ray, Glenn, | | What particular piece of "select" code are you two talking about??? | | ============================================================================ | | Wed, May 18, 2005 - 7:27:42 pm | | ray> | Joe, | | It's the bug fix for the crash that happens (happened!) if you | move the mouse off a pulldown menu. Seems the HTML tag involved | is called 'SELECT' ... but what do I know. See Glenn's fix in | MouseOn(), mouseon() in the old code, ver 1.83 or better. | | BTW, Joe, I now have an Australian Nephew in law, so I guess we're | related ;-) Can you give me a few choice Ausie insults that I might | use on him? The kidding sort, for teasing him. | | | | ============================================================================ | | Wed, May 18, 2005 - 7:31:21 pm | | glennmcc> | Hi Joe, | | This be the code in question :) | | --- in guitick.c ---- | | --- near the top --- | | #include "html.h"//!!glennmcc: Feb 18, 2005 -- for 'Select test' in mouseon( | | --- further down in the mouseon() function itself --- | | void mouseon(void) | { Š| if(!global_nomouse) | //!!glennmcc: Feb 15, 2005 -- fix pointer colors and 'stuck pointer' | { | //!!glennmcc: added 'Select test' Feb 18, 2005 | //calling onmouse(0) while in a | x_yncurs(1,mousex,mousey,15); | } | //!!glennmcc: end | } | | _____________ | | ============================================================================ | | Wed, May 18, 2005 - 7:32:33 pm | | glennmcc> | | | Hi Ray :) | | | | ============================================================================ | | Wed, May 18, 2005 - 8:43:38 pm | | ray> | Oh, I don't think mine will work in your code, basicaly I've moved | all mouse color work into onmouse() itself so MouseOn does nothing | but filter the SELECT tag, as you figured out, then it calls OnMouse(), | that's all. In fact, if I can figure out where this bad SELECT | problem comes from, I'll remove MouseOn entirely and just call | OnMouse directly. Besides MouseOn relies on the return value of | OnMouse to set one of two colors, but right now I'm using four colors: | Link, Neutral, Text, and Menu -- yes I know, it's just BS but | I'm having fun. If you're intested I'll post my OnMouse(). | | | | ============================================================================ | | Thu, May 19, 2005 - 1:48:35 am | | glennmcc> | DUDE......... Of course I'm interested. :) Š| | As for having fun....... I'm with you on that one. | | Why else are we doing this if not to have fun ? | | ============================================================================ | | Thu, May 19, 2005 - 5:21:34 pm | | ray> | OK, it's hardly finished work but I'll upload it to AD, I think | it's adaptable to the old code, just look for the new macro MOUSE. | | | ============================================================================ | | Thu, May 19, 2005 - 8:25:19 pm | | ray> | Ha, found the offending calls: If you want, you can move your filter | from MouseOn() into HideHighlight() and ShowHighlight() ... just filter | the call to MouseOn(). That's not your style, but it will be faster, | since the filter is only checked when needed. | | | | ============================================================================ | | Fri, May 20, 2005 - 9:49:20 pm | | ray> | Think I've got the *real* fix, or getting very close anyway -- no | need for filters. | | | ============================================================================ | | Sat, May 21, 2005 - 5:23:34 am | | Joe> | Hi Glenn, | | Thanks for the code above ... Yes, it does look like a slight | kludge, but nothing worth worrying about. OTOH, tracking down | the real culprit is hours of fun and helps to gain knowledge | of the code, right Ray? | | | ============================================================================ | | Sat, May 21, 2005 - 5:37:11 am | | Joe> | Hi Ray, | Š| Thanks for the uploaded code at AD, I'll study it a bit later (my | 'puter has been having lock-up problems these past couple of days, | just found and fixed the problem - an electo in the power supply | was starting to "go"). | | BTW, you might like to post an explanatory note at AD to accompany | your posting, there's 33 subscribers there, but I don't think many | of them read the "Board" (it doesn't forward them any postings by | e-mail;-). Also, if I can make a request, it would be best to try | to remember to include a Subject heading with all postings, it makes | threads easier to follow or to look up in the future (TIA:-). | | | | ============================================================================ | | Sat, May 21, 2005 - 5:54:07 am | | Joe> | [OT] | | Ray, | | I've been thinking about "insults" for you (for your nephew-in-law). | I'm probably not a good source for these, since I don't generally | engage in this practice (sure, I kid around with friends I know | well, but it's very specifically targeted and doesn't really involve | "insults" in any case). However, some guidelines : | | Firstly, don't start trying such "insults" until you know he won't | take them seriously! You may refer to him as a Kiwi; imply he's a | convict or of convict stock; if he's from NSW, refer to him as a | Victorian (or vice versa); find out what football team he follows | and kid him when they lose or their arch-rival is doing well. Hope | that's some help. If you have a particular "insult" in mind and | want an australian flavour to it, just ask (as long as you don't | blame me if it all goes wrong!;-). | | | ============================================================================ | | Sat, May 21, 2005 - 3:13:19 pm | | glennmcc> | Cludge ???? | | What make THIS a cludge ???? | | #include "html.h" | | | | if(activeatom.data1!=SELECT) | | Š| | else | x_yncurs(1,mousex,mousey,15); | | | Thats the whole fix..... 4 lines and nothing more. | | ============================================================================ | | Sat, May 21, 2005 - 3:33:23 pm | | ray> | Joe, | | Roger all that. I'll try the 'state' thing, he's from Sydney so | that's Victoria IIRC so I'll call him a 'bloody Queenslander' or | something, thanks :-) | | Glenn, | | Tisk tisk tisk, you don't know a kludge when you see one? | Sure it fixes it, but at the cost of a filter that is used | 100's of times when it isn't needed just to catch one as yet | undiscovered problem. I've almost found the exact cause of the | problem. And the right fix looks like it will remove lines, not | add them. | | | | ============================================================================ | | Sat, May 21, 2005 - 7:45:02 pm | | ray> | Well, it looks good. Code to AD tommorow. | | | ============================================================================ | | Sun, May 22, 2005 - 1:45:10 am | | gregy> | Joe, Ray, | | If Ray was from US (instead of that strange land just to the north), | since his nephew-in-law had married into a yankee family...he could | call him a "bloody septic tank"...and insult him royaly....but...since | Ray's a Canuck, heck....that's no option...... | | | ============================================================================ | | Sun, May 22, 2005 - 10:30:20 am | | ray> Š| Ha, I love that, that insult should be universalized. | | | ============================================================================ | | Sun, May 22, 2005 - 11:29:43 pm | | Joe> | Hi Glenn, | | Hey, don't fret ... a simple kludge beats an unresolved | problem any day of the week or year!!! :-) | | As I said before, it's nothing to worry about. Ray seems | keen to track this one down to its root cause, so lets see | what turns up ... | | | ============================================================================ | | Sun, May 22, 2005 - 11:33:36 pm | | Joe> | [OT] | | Very funny, Ray! | | ============================================================================ | | Sun, May 22, 2005 - 11:51:00 pm | | ray> | Oh ya, I should post that. To be brief, the ultimate cause is | in ... (nuts can't remember the func name) ... SelectSwitch() ?? | ... where there is no guard against running off the menu downwards. | Basicaly, SelectSwitch calls MouseOn, calls OnMouse, which tries to | draw a highlight where there is nothing to highlight which then | returns a bogus value back to SelectSwitch which ends up in a loop. | | You can break the cycle in MouseOn, but that results in the mouse | colors misbehaving (check 1.83) so I put a guard right in SelectSwitch | which notifies OnMouse not to try to highlight a nonexistant button. | | Clear as mud? I'll post it manana. | | | | ============================================================================ | | Sun, May 22, 2005 - 11:55:17 pm | | ray> | BTW, Joe, I wrote some vastly simpler dialer code that works fine for | me, (I figure I've sort of built in lspppcfg right into core). | It cuts out minitrue altogether, and just uses lsppp for everything. Š| Problem is it doesn't work for Vladimir. If he uses lsppp as his | dialer, he can't go on to ppp. Do you have any idea why that might | be? I know you mentioned that there were problems with lsppp as | a dialer, maybe this proves that that problem is still with us. | Comments? | | ============================================================================ | | Mon, May 23, 2005 - 1:37:36 am | | Joe> | Hi Ray, | | Miniterm, Miniture, it's all the same, no? ;-) | | Anyway, re LSPPP dialling - IIRC it works for me, although | it's been a looong time since I've tried and that was probably | with a defunct ISP anyway. It's one or two other users that | have complained about it (being unable to dial and connect | using just LSPPP) to me over the years, which has made me | stick with Michael's Miniterm. I am currently migrating to a | new PC at work, once that's done and things are back to normal, | I'll see if I've archived any of those e-mails about this issue. | Since I couldn't reproduce the problem at the time, it was not | practical for me to try to figure out the cause or a solution ... | | ============================================================================ | | Mon, May 23, 2005 - 8:57:59 am | | glennmcc> | Which version of lsppp is Vladimir using. | | There was one particular version in-which the dialer _was_ 'broken'. | (I don't remember the v#) | | Anyways... the bug is now fixed and the dialer works perfectly. | | ---- my current ppp.log pasted-in via F3 --- | | LADSoft PPP Packet Driver Ver 1.0 Copyright (c) LADSoft | Dialing...Connecting to network...Authenticating...Negotiating IP linkage... | Connection open | Installed packet driver handler at vector 0x60 | PPPD status: IP address set to 4.252.90.232 | ____________________________________________________________________________ | | | ============================================================================ | | Mon, May 23, 2005 - 9:06:00 am | | glennmcc> | --- pppdrc.cfg --- | 115200 Š| irq 4 | base 0xb800 | modem | crtscts | asyncmap 0 | namsrv | namsrv | user "glennmcc@cisnet.com" | passwd "*****" | ____________________ | | As you see... | no comport# (I have lsppp setup to use the base port address) | no DNS#s (lsppp gets them from the server) | | ---- lsppp.cfg --- | /M:atf50 | /x | /i:4 | /b:bx800 | /d:330-423-4005 | /B:115200 | /V:60 | /U:glennmcc@cisnet.com | /P:***** | _________________ | | --- my current environment --- | OS=OPENDOS | VER=7 | ARA=ON | PATH=\;C:\1BATCH;C:\1COMMON;C:\;C:\OPENDOS;C:\1ZIP;C:\QEMM;c:\borlandc\bin;c | OPENDOSCFG=C:\OPENDOS | BLASTER=A220 I5 D1 H7 T4 | MYIP=4.252.90.232 | REMIP=209.247.21.177 | NETMASK=0.0.0.0 | PEERMRU=1500 | DNS1=216.151.83.45 | DNS2=216.151.83.55 | COMSPEC=l:\command.com | TEMP=l:\temp | PROMPT=$d $b $t$_[OPENDOS 7.01] $P$G | ______________________________ | | ============================================================================ | | Mon, May 23, 2005 - 9:25:51 am | | ray> | Glenn, | | V says he's tried them all and even coresponded with the author | on the problem, so I don't think that's it. I like my new dialer | code, it essentialy formalizes what you are doing above, it just Š| expects you to put everything in lsppp.cfg or on the one 'connection' | line in arachne.cfg, but it doesn't work for Vladimir. No big deal, | the old code works fine for him, but I am very curious as to why | he needs to use miniterm as his dialer (which then requires all those | old keywords which could otherwise be removed). I'll show ya. | | | ============================================================================ | | Mon, May 23, 2005 - 9:31:46 am | | ray> | | | // RAY: 04-04-12: Moved from 'adial.c' only used here. | // Create contents of $roura$.bat for dialing. | // RAY: 05-05-05: After .c33, start tinkering. | // No use of 'port', must specify base and irq ... life is tough. | // No use of cBREAK. | // No terminal. | // dospppd always true, removed. | // No need for hangup string. Should be a test to avoid calling dialer when | online | // Moved 'unlink' calls to end. | // Removed errorlevel " /?" and skip lines from buf. | // Removed 'terminal'. | // Rework buffer write into single sprintf. | // Forget pppdrc.cfg, seems not needed for lsppp. | | static char *near ArachneDial(void) | { | #ifdef EX1 | { | char buf[5 * IE_MAXLEN]; // Buffer for command sent to $roura$.bat | | ProcessForm(0, IE_NULL); // update Arachne.Cfg | | // NB 'dialer' is expendable, but 'connection' is used in guievent.c and | // atcpip.c for various things as well. | if (!sUI.connection) | { | puts("\nSORRY, YOU HAVE NO 'Connection' SET IN ARACHNE.CFG"); | delay(4000); | return (""); | } | sprintf(buf, "%Ns\n" | "@if exist IP-UP.BAT call IP-UP.BAT\n" | "@echo PPPD status: %Ns %MYIP%>>PPP.LOG\n" | ,sUI.connection, sUI.ip_grab); // "IP address set to" ... could be cMESSAGE. | | // There's gota be an easier way! | if ((!strcmpi(sUI.dialpage, p->htmlframe[p->activeframe].sCacheItem.URL) | || strstr(p->htmlframe[p->activeframe].sCacheItem.URL, "err_")) | && sArachne.scriptline == 0) | Š| strcat(buf, "@arachne.bat -o\n"); // Ends up at homepage, or shows error | else // see INIT.C c. L 630 and 735. | strcat(buf, "@arachne.bat -r\n"); // Reload previous page. | | // A TYPICAL 'buf' AS PRODUCED BY THE ORIGINAL CODE: | /* | @lh lsppp.exe /d:6046788350 /M:atf50 /n:2 /B:115200 /U:randrews ... | @if errorlevel 1 goto skip // What does this do? | @lsppp.exe /?>NUL // What does this do???? Help screen??? | :skip | @if exist IP-UP.BAT call IP-UP.BAT | @echo PPPD status: IP address set to %MYIP%>>PPP.LOG | @arachne.bat -r | */ | | // Buf2File("temp.txt", buf); | Outsf(cDIAL, cESC); | | // unlink("IP-UP.BAT"); // Used by lsppp.exe or epppd.exe to set env. variab | // unlink("PPP.LOG"); // Must delete to reset online timer but this | // is done in arachne.bat | return buf; | } | #else | { | ... old code ... | | | ============================================================================ | | Mon, May 23, 2005 - 4:51:57 pm | | glennmcc> | And you called my code a kludge ???? | | ============================================================================ | | Mon, May 23, 2005 - 5:36:00 pm | | glennmcc> | Just back from looking-over you most recent post to AD (#1726) | | Hmmmm.... | | Many lines of code spread-out over several SRC files. | | versus | | 4 lines of code located in 1 SRC file. | | == no-brainer | | ============================================================================ | | Mon, May 23, 2005 - 7:07:05 pm Š| | Joe> | Hi all, | | Glenn, if it makes you feel any better, and to put things in | perspective, I consider the whole LSPPPCFG package to be a kludge. | | | I'll have to look at Ray's code (when I have time!), before I can | give an opinion on it ... | | | | ============================================================================ | | Mon, May 23, 2005 - 8:05:03 pm | | glennmcc> | Thanks for that Joe. | | It's nice to hear that someone else around here actually _likes_ a good klud | | | IMO, | Actually, when we really dig into her.... Arachne herself is the | culmination of hundreds of 'good kludges' combined together to make a | very useful and indispensible 'finished product'. | | ============================================================================ | | Mon, May 23, 2005 - 11:04:51 pm | | ray> | Glenn, | | I knew you'd react that way. Never mind, you fix does have the | virtue of beeing simpler to look at, but mine has the advantage | of being faster and actualy fixing the problem. Beware fixes that | create bottlenecks. | | | ============================================================================ | | Tue, May 24, 2005 - 4:35:43 pm | | glennmcc> | You know what ? | | I am about sick-and-tired of this bickering that goes on between the two of | And I now intend to put a stop to it. | | I would like to ask a favor of you, if I may. | | I'll zip-up my current SRC code changes made between v1.83 and the | current one I'm working on. (v1.85) Š| | I'll make one last release in a few months using my version of the code | which is of-coarse the original SRC with the various changes made over | the last few years. | | Now for the favor.... If you could please 'meld' my 1.83-->1.85 changes | into your current itteration of the SRC and then zip-up the whole thing | to send back to me. | | My itteration of the SRC will then be zipped-up and set-aside. | | It's high-time I jumped onto your band wagon and started using the cleaned c | | | What you think ? ;-) | | | ============================================================================ | | Tue, May 24, 2005 - 9:04:13 pm | | ray> | Glenn, | | Welcome aboard! But as I said some time ago, I've realized that | it is a good thing to keep both lines going as long as you've got the | patience to do it. But if you want to develop in the cleaned code, | that's great. As far as melding goes, that's basicaly standard | practice anyway although I have left a few things out. However, if | you're on board we can make an integrated base with any code | disagreaments handled via #defines. If Joe accepts base 5x then we | can add a define for him to should there be anything of his that | the others of us don't like. | | ============================================================================ | | Tue, May 24, 2005 - 9:09:00 pm | | ray> | Oh, one problem, as you know, up to about base3x I made an effort to | do things by consensus, but after I more or less figured you'd not | accept my code, I've started to do a lot of things 'my way', however | I don't think there's too much that's objectionable. But I will need | to return to consensus mode, esp. if Joe 'joins'. Right now I'm | working on B6 -- another end to end edit to see what I can see. | No big changes really. | | | ============================================================================ | | Tue, May 24, 2005 - 9:11:52 pm | | ray> | ... but it will need to be shaken down. Give me a month to finish | this edit, then incorporate your and Joe's changes then she should Š| be ready to shake down. | | BTW Joe, | | Vladimir reports that your code handles trackballs properly while | mine and Glenn's don't. Did you kill some bug or other there? | | | | | ============================================================================ | | Tue, May 24, 2005 - 9:30:21 pm | | glennmcc> | One month, you say ? | That's kewl by me. :) | | As I mentioned... in a few months I'll release v1.85 compiled using | my itteration of the SRC | | After that.... I'll move to your itteration. :) | | My tentative release date is set for the 3yr anniversary of v1.71;UE01 | | v1.71;UE01 .......... Release date: Oct 30, 2002 | v1.85;GPL .......... Release date: Oct 30, 2005 | | What say all ;-) | | ============================================================================ | | Tue, May 24, 2005 - 10:34:25 pm | | ray> | That's great, I might even get to enjoy the Summer :-) | ... Make it *two* months < g g g > | | BTW, just today I think I've finaly (well, this is about the 5th | 'finaly') come up with a decent fix for the bloody 'go back' | thing. | | Oh, Glenn, if you're coming over to cleaned code, just say the word | and I'll send you my latest work in progress -- may as well start taking | a look, so we can haggle over any issues you might have or bugs you | might find. And speaking of haggling. I'm teatering on the brink | of moving the whole project to two space indentation, what say? | IIRC that was Joe's preference too. | | | ============================================================================ | | Tue, May 24, 2005 - 11:38:49 pm | | glennmcc> Š| Two space or not two space...... that is the question. ;-) | | Two space it is then. :) | | | ============================================================================ | | Wed, May 25, 2005 - 9:45:31 am | | ray> | Joe? | | | ============================================================================ | | Thu, May 26, 2005 - 7:14:23 am | | Joe> | Hi all ... | | Bickering? Not much I've seen here would constitute bickering, IMHO. | Cross-fertilization of ideas is beneficial for the development of | Arachne, not a hinderance. I think we could do with even more input | from other developers, it's a pity we don't have Bernie and Michal | and whoever else also giving us a broader perspective. Anyway, enough | of philosophy . | | As for the "clean code" branch(s), Glenn's renewed interest means I | need to find more time to complete the 1.83C1 code, which of course | is based on Ray's Base 3 (3C?) code. This may be a suitable "clean | code" candidate, since this was a consensus branch. What do you think? | | | | ============================================================================ | | Thu, May 26, 2005 - 7:18:15 am | | Joe> | Re trackballs, I dunno. I implemented Laaca's code fix, other than | that, I can't think what other subtle differences in the 1.83J1 | code would account for Vladimir's observations. | | | | ============================================================================ | | Thu, May 26, 2005 - 7:27:07 am | | Joe> | Re two-space indentation, certainly, this is what I would favour for | Arachne. For practical (pragmatic;-) reasons, complex code must use | the minimum indentation that is still reasonably visible. Hence, we | arrive at the two-space indentation (since single-spacing has poor | visibility), arguably the most widely adopted indentation for complex Š| coding projects. | | | ============================================================================ | | Thu, May 26, 2005 - 5:33:47 pm | | ray> | Joe, | | The 'bickering' comment was more dry humor :-) | | Looking forward to your 3C-line updates. I don't know if Glenn | will like that more than 5-level code. The latter has a lot | more work done on it. I'd like your opinion if you have time to | give it a look. The trackball thing is strange, we'll figure it out. | Ok, it's a go for two space indentation. Three space was usefull while | I was sorting out the indentation messes way back when, but it's | served its purpose. | | ============================================================================ | | Fri, May 27, 2005 - 3:26:58 am | | Joe> | Hi Ray, etc. | | OK about the "dry humour" - the lack of and/or ;-) in situ | had me thinking Glenn was serious ... | | About Base 5, doesn't that exclude POSIX and Linux code??? | | | ============================================================================ | | Fri, May 27, 2005 - 2:08:03 pm | | ray> | Joe, | | Yup, it does, but it looks like my predictions vis. nothing much | happening on that front are proving correct. But I stick to my | promise on that -- if we ever do get a decent linux project going | I will integrate any code we need, but I continue to believe that | a new linux project will basicaly be a start from scratch affair and | very likely will procede with a different compiler which will almost | certainly mean a code fork in any case, so there's little to | be gained by keeping that stuff in our *working* DOS build. | | | | ============================================================================ | | Fri, May 27, 2005 - 2:09:51 pm | Š| ray> | Oh, forgot: Joe, I neglected to remember that you aren't on the | arachne4dos list -- 'dry humor' has been a big topic lattely :-) | | | ============================================================================ | | Fri, May 27, 2005 - 4:00:06 pm | | Joe> | Hi Ray, etc. | | Yup, if I had any spare time on my hands, I'd probably indulge in | chat at A4D, however, such is not the case unfortunately. Other | lists I've belonged to but had to stop participating in for the | same reason : | | - arachne @ arachne,cz (joined for just 24h, had tons of messages | in that time, none of which were relevant ... somwhat like A4D) | - interfacing @ egroups (now yahoo) | - chipdir | - gnu pascal | - survpc | - a couple of others that I can't recall at the moment | | :-( | | | | ============================================================================ | | Fri, May 27, 2005 - 4:51:07 pm | URL | | glennmcc> | Ray, | | Would you care to make a bit of a waiger on that ??? | | http://www.cisnet.com/glennmcc/my-stuff/aralinux.zip | | ============================================================================ | | Fri, May 27, 2005 - 9:05:10 pm | | ray> | Ah, the URL isn't working right now, but I presume it's about the | linux thing? ... sure I'll take a bet on that but only on | condition that I loose. | | ============================================================================ | | Fri, May 27, 2005 - 11:04:51 pm | | glennmcc> Š| hehehe ;-) | | | ============================================================================ | | Fri, May 27, 2005 - 11:06:08 pm | | glennmcc> | BTW, | | I'm here right now using arachne-svgalib v1.66b on SlackWare Linux v7.1 :) | | ============================================================================ | | Fri, May 27, 2005 - 11:06:56 pm | | glennmcc> | My 'fresh compile' that-is... not the original :)) | | ============================================================================ | | Sat, May 28, 2005 - 4:57:08 pm | | ray> | Most cool :-) So we have hot linux code. She'll be a long time in | bringing her up to speed with insight, etc, but this is the future. | Nuts, why can't we find some linux guru to work on this? Linux | people can be as excentric as DOS people and there's more of them, | there's gota be somone out there who wants to bring 1.66 up to | speed. Once she's buildable with 3c code and all the crap has been | shaken out, it should be just a few months work to incorporate any | POSIX stuff into the lattest cleaned code, then we're off to the | races. | | | | ============================================================================ | | Sat, May 28, 2005 - 5:09:12 pm | | glennmcc> | I have not looked closely... | | Does 3c have all of the POSIX already removed ? | | ============================================================================ | | Sat, May 28, 2005 - 7:26:51 pm | | ray> | Nope, it's all there in 3c. I had it commented in 3a but at Joe's | request, it's all colorful in 3c. Of course not one line of it | has been debugged from any errors introduced durring the base1 | meatgrinder stage. If you or anyone else were going to look at that, | start with B1 and move forward from there. Š| | | ============================================================================ | | Sun, Jun 05, 2005 - 9:36:51 pm | | ray> | BBBBbbbut I'm confused: I thought the reason for the hang with those | bad emails was that they stopped downloading before they were finished. | If that were the case, then 'locallength' would never reach | zero anyway, no? Assuming that the size of the file is reported | corectly by the server, it would seem to me that the hang would happen | when the bad characters were encountered and before locallength | was legitimately at zero. | | | | ============================================================================ | | Sun, Jun 05, 2005 - 9:38:28 pm | | ray> | ... unless, of course, the problem was precicely that the reported | size is wrong ... | | ============================================================================ | | Sun, Jun 05, 2005 - 9:52:16 pm | | ray> | ... and I don't have a firm grasp of what's going on, but it seems | to me that all that 'done' stuff looks pretty greasy. What would | happen if we just relied on 'locallength' to terminmate a download | instead of all those strange tests for various dots and '\n' etc?? | I have all that stuff commented right now. lets see if I can send | myself and email ... | | ============================================================================ | | Sun, Jun 05, 2005 - 9:58:47 pm | | ray> | Yup, I can send myself an email. So that stuff must be | bug handling? Or bad email handling? | | | | ============================================================================ | | Mon, Jun 06, 2005 - 7:11:45 am | | ray> | Oppps, this morning's email wouldn't download with that stuff commented. | | Š| ============================================================================ | | Mon, Jun 06, 2005 - 2:15:31 pm | | ray> | But, today I needed to comment Glenn's 'suplementary fix' to download | my mail. Far to early to say if this is conclusive tho. | | | ============================================================================ | | Mon, Jun 06, 2005 - 6:21:32 pm | | glennmcc> | Nope.... they did not stop _before_ the end was reached. | | The freeze-up happened at the exact end of the file. | | That's why the fix works. | | As the code was before..... it did not proceed-on to the next file | when the end of the current file was reached. | | It would only proceed-on _if_ it encountered any of those conditions | which set done=1 | | If it did not encouter any of them......... freeze-up. :( | | We now have double protection because | | 1) the do/while now contains 2 conditions..... while(!done && locallength>0) | 2) if(locallength<1) done=1; //double protection ;-) | | | ============================================================================ | | Mon, Jun 06, 2005 - 9:57:16 pm | | ray> | Well, I don't understand this stuff anyway so pardon my questions, | but what does 'done' do? Seems to me that an email is done | when the full length of it is downloaded, but my trial commenting out | of those tests didn't work so those wierd character tests are | obviously important, but when would they have effect even though | the length of the file is not yet reached? In other words, what | tests are needed *besides* 'locallength == 0' (end of message)? | | | | ============================================================================ | | Tue, Jun 07, 2005 - 7:47:00 pm | | glennmcc> | In-order to answer that question, I would need to se axactly what it was Š| that you commented-out. | | | As to the rest of it........ | | Everthing is being done inside of a do/while 'block'. | | do | { | query server for file length; | | set 'locallength' to returned value; | | download this message one-line-at-a-time; | | decrement 'locallength' by the number of characters in this line; | | check to see if any of the conditions have been met which will | set 'done=1'; | | } | | while(!done) | | __________________________________ | | The do/while continues to be executed over-and-over again | grabbing one line at a time until one of those conditions is met which | sets the variable 'done' to 1 | | There was only circumstance which could end the do/while .... the | setting of 'done=1' | (no test to end the do/while when the end of the file was reached) | | Therefore, if the end of the file was reach and none of the conditions | had been met for setting done=1............ freeze-up. :(((( | | Now the do/while will be ended on 2 different cicumstances instead of | just one. | | while(!done && locallength>0) | | Now when the end of the file is reached, locallength will be 0 | | Since it is no longer greater than zero, the do/while has now ended and | we are moved on to the next message....... no freeze-up. :))) | | | ============================================================================ | | Wed, Jun 08, 2005 - 4:23:31 pm | | ray> | Well that's clear enough, but I'm asking the 'opposite' question, | as it were, I see that the 'locallength' test is needed but I'm Š| wondering why the other tests are. IOW when would a message be | terminated before locallength == 0? | | | | ============================================================================ | | Wed, Jun 08, 2005 - 8:06:26 pm | | glennmcc> | My current compile is using only this and nothing more..... | | while(locallength>0) | | No other tests of any kind... and when I compile core.exe I get one | 'warning'............. done is assigned a value that is never used. | | That is due to the fact that I did not comment-out these 2 lines in popget.c | | char done,lastchar[2]; | | done=0; | | | With this fresh compile using only locallength>0 as the exclusive | 'test' for continueing the do/while | | Everthing is working perfectly. | | I still leave the older compile in-place which does not include the fix | for the 'freeze-up SPAMs' | | When a freeze-up happens, I switch to the new compile and try | downloading email again. | | In my 'first run' of email today.... | #11 of 27 was yet another of the 'freeze-up SPAMs' | | Upon downloading email a second time, when #11 was reached that froze-up | the old compile, the new compile prevented the freeze-up and message #12 | through #27 come down the pipe with no problems at-all. | | I have continued using the new compile with only while(locallength>0) | for several more 'email runs' this evening. | | Each and every time.... everything worked perfectly. | | My conclusion: | | while(locallength>0) is all we need. | | We can remove all other tests and we no-longer need the variable 'done' at-a | | ============================================================================ | Š| Wed, Jun 08, 2005 - 9:51:26 pm | | ray> | Hmmm, I tried that, and I kept getting aborts, maybe I did something | wrong, I'll try it again, it seems like it should work, no | weird end of message tests etc. If I do get aborts, it should be | easy enough to figure out why. | | | ============================================================================ | | Wed, Jun 08, 2005 - 10:27:56 pm | | glennmcc> | To watch exactly what's going on .... | | Uncomment those 2 lines where I indicated... | "the next 2 lines were used during my testing" | | Uncommening them will show the current value of locallength at the | bottom of the screen each time it gets decremented. | | When it reaches zero, the do/while will be broken-out-of and the next | waiting message will begin to download. | | | ============================================================================ | | Thu, Jun 09, 2005 - 8:57:08 pm | | ray> | Right, will do. | | | ============================================================================ | | Fri, Jun 10, 2005 - 2:22:50 am | | Joe> | Caution, I may not know what I'm talking about here ... | | 1. It sounds to me like the variable 'locallength' is a misnomer, it | sounds more like 'bytesleft'. | | 2. I suspect the other "done" conditions (which I haven't looked at, | see caveat;-) may be to handle problems due to communications errors. | They may not be evident unless you have bad line conditions with | your modem or whatever. | | 3. If speculation (2) is true, then you probably should have a new | condition that sets 'done' to true (1) when 'locallength' reaches 0, | rather than supplanting 'done'. | | ============================================================================ | Š| Fri, Jun 10, 2005 - 6:55:09 pm | | glennmcc> | Joe, | | caveat == correct | | 1 == partly correct | | locallength == filesize of local file nnnnnnnn.cnm | locallength == decremented by current line length so that it is also | being used as the indicator of bytesleft to download. | | both 2 and 3 == incorrect | | Those 'done' conditions were being used for 'special cases' where the | file being downloaded di not end in the 'normal manner' of LF DOT | | Therefore it changed those last 2 bytes so that they would conform to | the accepted standard for emails. | | And then it set done=1 so that the downloading of the current file would | end and the downloading of the next file could begin. | | _____ | | | My current 'testing code' does not worry about 'non compliant' emails. | | It simply ends the current download and begins the next as soon as the | end of the file is reached. (regardless of what characters appear at the | end of this current file) | | No changes are made... the file is saved to disk as nnnnnnnn.cnm | identical to how it was downloaded. | | | ============================================================================ | | Fri, Jun 10, 2005 - 8:31:10 pm | | ray> | Glenn, | | I've been downloading with the 'done' stuff comented out, and I get | aborts left right and center. Looking at pop3.log, I see that | any message ending without the dot at the end is the culprit. As for | locallength, its value ocasionaly dips into the negative 100's, which | is a bit disconcerting if that is supposed to be the exact lenght of | the file to be downloaded. | | ============================================================================ | | Fri, Jun 10, 2005 - 8:34:19 pm | Š| ray> | What I can't understand is why it is when the 'done' stuff is | NOT active that I get these hangs with messages that DON'T have | the bleeding dot when that ain't being tested for anyway. I'm glad | this is your project ;-) | | | ============================================================================ | | Fri, Jun 10, 2005 - 8:36:24 pm | | ray> | Joe, | | Reduction in speed confirmed in your latest as well. Interesting that | the change in speed apears to be exactly 2/3, this leads me to | suspect that your .lib isn't 'broken' but some paramater or other | has been altered. Just a guess of course. | | | | ============================================================================ | | Fri, Jun 10, 2005 - 8:50:02 pm | | glennmcc> | Try these URLs for speed testing. | | Since accessing the files via the IP address instead of the domain | name 'bypasses the DNS... this will eliminate and possibility of DNS | being the cause. | | http://68.22.80.45/glennmcc/images/linux_kde_netscape.gif | http://68.22.80.45/glennmcc/images/slck10_1.jpg | | ============================================================================ | | Fri, Jun 10, 2005 - 9:08:10 pm | | glennmcc> | Just now completed 5 consecutive downloads with each of my 2 cores. | (the 2 testing cores were built with only one difference between them) | | 159,232 6-06-05 8:35p e:\wattcp\wat1104\lib\glennmcc\wattcplg.lib | (built using port 997 changed to port 0 in UDP_DOM.C | but using the original PCTCP.C instead of Joe's altered version... | ...no other difference at-all) | (average through-put == 5300 bytes per sec) | | 159,232 4-22-05 1:06a e:\wattcp\wat1104\lib\joe\wattcplg.lib | (average through-put == 3500 bytes per sec) | | | ============================================================================ | Š| Fri, Jun 10, 2005 - 9:31:24 pm | | glennmcc> | Ray, | | Could you please zip-up your altered POPGET.C and send it to me ? | | I'll see if I can find the cause for those aborted downloads you are getting | | | ============================================================================ | | Fri, Jun 10, 2005 - 11:09:23 pm | | gregy> | Glenn, was that what you were looking for...?? I mean the download | speed differences I recorded...??? | | | | ============================================================================ | | Sat, Jun 11, 2005 - 5:34:59 am | | Joe> | Hi Glenn, etc. | | Thanks for the explanations re e-mail downloads ... as you say, | caveat == correct ... | | | ============================================================================ | | Sat, Jun 11, 2005 - 5:44:28 am | | Joe> | Hi again, | | Interesting about the speed issue ... I don't see it at this end, | I'm using LSPPP 1.01, but I doubt that makes a difference. | | Does using IP addresses instead of domain names make any difference, | as you've suggested above as a test? | | Also, does the inital download take a little while to start, or | does the whole download proceed more slowly? Arachne would only | be doing a DNS look-up to start a download, thereafter the data | should come in without any further DNS interaction. | | If these two clues point to DNS, I'll have another look into | this stuff as a matter of priority. | | | | ============================================================================ Š| | Sat, Jun 11, 2005 - 2:55:12 pm | | glennmcc> | Nope... it makes no difference at-all. | | The same problem exists with both of these downloads... | | http://www.cisnet.com/glennmcc/images/linux_kde_netscape.gif | http://68.22.80.45/glennmcc/images/linux_kde_netscape.gif | | | Hence since when using the IP address instead of the domain name, | no DNS is ever contacted to ask for the IP address of that domain name.... | we have proven that the problem is not with DNS access. | | ============================================================================ | | Sat, Jun 11, 2005 - 3:01:51 pm | | glennmcc> | This is the only difference between the 2 builds of WATTCPLG.LIB | | | This build .... | 159,232 6-06-05 8:35p e:\wattcp\wat1104\lib\glennmcc\wattcplg.lib | | Uses this file.... | 94,724 5-23-01 1:17p e:\wattcp\wat1104\src\original\pctcp.c | | This build..... | 159,232 4-22-05 1:06a e:\wattcp\wat1104\lib\joe\wattcplg.lib | | Uses this file..... | 95,051 4-21-05 11:20p e:\wattcp\wat1104\src\jds\pctcp.c | | | That's it...... no other differences of any kind. | | ============================================================================ | | Sat, Jun 11, 2005 - 3:04:08 pm | | glennmcc> | Oh yeah, | | Both builds use this file...... | 18,836 4-20-05 11:00p e:\wattcp\wat1104\src\jds\udp_dom.c | | | ============================================================================ | | Sat, Jun 11, 2005 - 3:05:42 pm | | glennmcc> Š| Gregy, | | Yep, that's what I was looking for. | | Thank you very much for those test results. | | ============================================================================ | | Sat, Jun 11, 2005 - 8:29:25 pm | URL | | glennmcc> | Son-of-gun it worked. :)))) | | I had a little bit of a hunch. | | Joe, you mentioned that you had some difficulty in getting the new LIB | to compile. (just as I had mentioned that I had) | | In looking inside of the original LIB from Nov 2004, | I found that it had been compiled with Borland C installed in C:\bc31 | | Using this bat... | | --- pathwat.bat ---- | | @echo off | c: | cd \ | rendir \borlandc \bc31 | SET PATH=\;C:\1BATCH;C:\1COMMON;C:\;C:\OPENDOS;C:\1ZIP;C:\QEMM;c:\bc31\bin;c | e: | cd \wattcp\wat1104\src | ____________________ | | I renamed my dir and reset my path as needed and rebuilt the LIB using | both of your changed SRC files. | | | The fresh copile of core.exe I am using right now uses the fresh build | of wattcplg.lib at the URL above. | | Average through-put on the test files.... 5300 CPS | | | ============================================================================ | | Sat, Jun 11, 2005 - 8:41:43 pm | | glennmcc> | And now comparing the original, my fresh build and your build..... | | Turbo Assembler Version 3.1 (original) | Turbo Assembler Version 3.1 (fresh build) | Turbo Assembler Version 3.2 (your build) Š| ---------------------------^ | | Now the question is...... why would the newer version of Turbo Assembler | have any effect on this ??? | | | | ============================================================================ | | Sun, Jun 12, 2005 - 9:29:11 am | | ray> | Glenn, | | Ok I'll send my popget. And that's good slewthing vis the .lib | thing, otherwise we'd have quite a puzzle to figure out. | | | ============================================================================ | | Sun, Jun 12, 2005 - 9:56:26 am | | ray> | Confirm new lib works at full speed :-) | | ============================================================================ | | Sun, Jun 12, 2005 - 9:56:57 am | | ray> | 0 | 1 | 2 | 3 | | | | ============================================================================ | | Sun, Jun 12, 2005 - 2:34:45 pm | | glennmcc> | Thank you Ray. | | popget.c edited and sent-back via email. | (also uploaded to Yahoo so that I could test uploading with Lynx 386 v2.85.1 | | ============================================================================ | | Sun, Jun 12, 2005 - 3:49:14 pm | | ray> | Ok, will test this evening. | | ============================================================================ Š| | Mon, Jun 13, 2005 - 1:01:37 am | | gregy> | NOTE - Re: Scott's post to coollist, about him having his compile setup | set up, let him get up and running and see how he integrates into the | development group, then (but I suggest not right away) give him this | URL so he can interact with you guys here. | | | | ============================================================================ | | Mon, Jun 13, 2005 - 1:07:43 am | | gregy> | AND - I see Scott's already signed himself up to the | arachnedevelopment@yahoo.com list... | .. | | ============================================================================ | | Mon, Jun 13, 2005 - 10:00:29 am | | ray> | The more the merrier. | | | ============================================================================ | | Mon, Jun 13, 2005 - 3:54:13 pm | | gregy> | res, definitely, and I think Scott will maybe be an asset. But, you | want to see how he integrates in, before introducing him to this place. | | | | ============================================================================ | | Mon, Jun 13, 2005 - 7:16:54 pm | | glennmcc> | It might also be a good idea to wait on introducing him to this place | til _after_ you have made the needed 'upgrades' so that pasted-in code | will display properly. | | ============================================================================ | | Mon, Jun 13, 2005 - 9:12:15 pm | | ray> | Yes! Gregy, you keep promising to fix that :-( | | Š| ============================================================================ | | Mon, Jun 13, 2005 - 10:43:12 pm | | Joe> | Hi Glenn, et al. | | Thanks for the feedback re wattcp library. I'm relieved it's not the | code changes that have made this slow-down effect for some (I don't | get this effect at this end). I started making the necessary changes | to keep a constant port number for DNS responses - it gets a bit | messy - I'll discard this effort now. | | It's bizarre that using TASM 3.2 may have been responsible, really | bizarre, given that the standard wattcp libary distribution is a | BC 4 compilation, so probably uses an even newer version of TASM. | | I think I'll do a fresh compilation and send this to Ray for some | quick testing - I just can't believe its a TASM version issue - I | am also astounded that things should still work OK, just more slowly, | you'd think the wattcp library either works or it doesn't, let alone | works a full speed for some, 2/3 speed for others. Arghhh!!! | | | ============================================================================ | | Tue, Jun 14, 2005 - 5:49:13 am | | Joe> | Fresh wattcplg.lib compilation - once again, initial compilation | results in an incomplete (slightly smaller) library, further | compilation is required to get a complete (full size) library. | | New, fresh, complete, compilation is the same size as the first | compilation of the same code. However, binary comparison shows | several dozen bytes have changed. Strange! | | Ray, I'll do a bit of testing on this fresh library compilation, | then if all seems well, I'll e-mail it to you tomorrow. I really | need to know if TASM 3.2 is the culprit for your speed differences | (I hate these unexplained/bizarre mysteries;-)! | | | | ============================================================================ | | Tue, Jun 14, 2005 - 10:55:26 am | | ray> | Joe, | | Ok, I'll expect that. Yes, we can't have any mysteries! Gotta | find out what's up. And I agree, a 2/3 slowdown seems the most | unlikely of results from some compilation goof up. | Š| | ============================================================================ | | Tue, Jun 14, 2005 - 5:57:39 pm | | ray> | Glenn, | | from the bottom of my pop3.log after another abort using 'locallength' | only to end the while(): | | Send operational and technical questions to :- | | developers list; arachne@freelists.org | | | [Sponsored by:] | ____________________________________________________________________________ | TheDELE 1 << WHAT DO YOU MAKE OF THAT? | newest lyrics on the Net! | Closing connection... | | | | ============================================================================ | | Tue, Jun 14, 2005 - 7:05:35 pm | | glennmcc> | That is the coomand that Arachne sends to the sever when the D/L has | completed. | | BRB | | | | ============================================================================ | | Tue, Jun 14, 2005 - 7:09:03 pm | | glennmcc> | Asking domain name server for mail.cisnet.com ... | +OK VOPmail POP3 Server 4.5.186.0 Ready <48627720.1118797679.323@cisnet.com> | USER glennmcc | PASS | +OK glennmcc's mailbox has 2 message(s) (1559 octets) | LIST 1 | +OK 1 778 | RETR 1 | +OK 778 octets | Return-path: | Received: from cisnet.com (unverified [4.252.95.184]) by mail.cisnet.com | (Vircom SMTPRS 4.5.186) with ESMTP id for | ; | Tue, 14 Jun 2005 21:07:51 -0400 Š| Message-ID: | From: "Glenn McCorkle" | Organization: Arachne Fan Club | Date: Tue, 14 Jun 2005 20:07:37 -0500 | X-Mailer: Arachne v1.85;UE08 | To: glennmcc@cisnet.com | Reply-To: glennmcc@cisnet.com | Subject: test #1... | MIME-Version: 1.0 | Content-type: text/plain; charset=ISO-8859-1 | Content-transfer-encoding: 7bit | | 1st test | | | -- | Glenn | http://www.delorie.com/listserv/mime/ | http://www.cisnet.com/glennmcc/ | http://www.law.cornell.edu/constitution/constitution.table.html | . | DELE 1 | +OK message 1 deleted | LIST 2 | +OK 2 781 | RETR 2 | +OK 781 octets | Return-path: | Received: from cisnet.com (unverified [4.252.95.184]) by mail.cisnet.com | (Vircom SMTPRS 4.5.186) with ESMTP id for | ; | Tue, 14 Jun 2005 21:07:52 -0400 | Message-ID: | From: "Glenn McCorkle" | Organization: Arachne Fan Club | Date: Tue, 14 Jun 2005 20:08:07 -0500 | X-Mailer: Arachne v1.85;UE08 | To: glennmcc@cisnet.com | Reply-To: glennmcc@cisnet.com | Subject: test #2... | MIME-Version: 1.0 | Content-type: text/plain; charset=ISO-8859-1 | Content-transfer-encoding: 7bit | | 2nd test... | | | -- | Glenn | http://www.delorie.com/listserv/mime/ | http://www.cisnet.com/glennmcc/ | http://www.law.cornell.edu/constitution/constitution.table.html | . | DELE 2 | +OK message 2 deleted Š| Closing connection... | +OK cisnet.com POP3 server signing off (mailbox empty) | | ============================================================================ | | Tue, Jun 14, 2005 - 7:13:15 pm | | glennmcc> | | Volume in drive C is MCCORKLE_01 | Directory of C:\ARACHNE\MAIL | | A2AF7176 CNM 779 6-14-05 8:08p | A2AF7177 CNM 782 6-14-05 8:08p | | As you can tell from the fact that each of the 2 files is one byte | larger than is indicated by pop3.log .... I have in-place the older | compile of core.exe which is still adding a LF to each file. | | BRB with the results of 2 more tests with the newer compile. | | | | | ============================================================================ | | Tue, Jun 14, 2005 - 7:16:44 pm | | glennmcc> | Asking domain name server for mail.cisnet.com ... | +OK VOPmail POP3 Server 4.5.186.0 Ready <48622064.1118798072.493@cisnet.com> | USER glennmcc | PASS | +OK glennmcc's mailbox has 2 message(s) (1562 octets) | LIST 1 | +OK 1 781 | RETR 1 | +OK 781 octets | Return-path: | Received: from cisnet.com (unverified [4.252.95.184]) by mail.cisnet.com | (Vircom SMTPRS 4.5.186) with ESMTP id for | ; | Tue, 14 Jun 2005 21:14:08 -0400 | Message-ID: | From: "Glenn McCorkle" | Organization: Arachne Fan Club | Date: Tue, 14 Jun 2005 20:14:28 -0500 | X-Mailer: Arachne v1.85;UE08 | To: glennmcc@cisnet.com | Reply-To: glennmcc@cisnet.com | Subject: test #3... | MIME-Version: 1.0 | Content-type: text/plain; charset=ISO-8859-1 | Content-transfer-encoding: 7bit | Š| 3rd test... | | | -- | Glenn | http://www.delorie.com/listserv/mime/ | http://www.cisnet.com/glennmcc/ | http://www.law.cornell.edu/constitution/constitution.table.html | . | DELE 1 | +OK message 1 deleted | LIST 2 | +OK 2 781 | RETR 2 | +OK 781 octets | Return-path: | Received: from cisnet.com (unverified [4.252.95.184]) by mail.cisnet.com | (Vircom SMTPRS 4.5.186) with ESMTP id for | ; | Tue, 14 Jun 2005 21:14:30 -0400 | Message-ID: | From: "Glenn McCorkle" | Organization: Arachne Fan Club | Date: Tue, 14 Jun 2005 20:14:49 -0500 | X-Mailer: Arachne v1.85;UE08 | To: glennmcc@cisnet.com | Reply-To: glennmcc@cisnet.com | Subject: test 4.... | MIME-Version: 1.0 | Content-type: text/plain; charset=ISO-8859-1 | Content-transfer-encoding: 7bit | | 4th test... | | | -- | Glenn | http://www.delorie.com/listserv/mime/ | http://www.cisnet.com/glennmcc/ | http://www.law.cornell.edu/constitution/constitution.table.html | . | DELE 2 | +OK message 2 deleted | Closing connection... | +OK cisnet.com POP3 server signing off (mailbox empty) | | | | | Volume in drive C is MCCORKLE_01 | Directory of C:\ARACHNE\MAIL | | A2AF72FF CNM 781 6-14-05 8:14p | A2AF7300 CNM 781 6-14-05 8:14p | Š| | Now they are the same size as indicated. (no longer adding the LF) | | ============================================================================ | | Tue, Jun 14, 2005 - 7:19:03 pm | | glennmcc> | Let's have a look at your entire pop3.log | | | | ============================================================================ | | Tue, Jun 14, 2005 - 7:24:02 pm | | glennmcc> | I just checked it. | | The point at-which the DELE command appears is 85 bytes from the actual | end of the file. | | It looks as-if your server is telling you a filesize which is 85 bytes | smaller than the actual figure. | | | ============================================================================ | | Tue, Jun 14, 2005 - 9:55:48 pm | | ray> | Hmmm, nice work, now why would they do that, the twits? | But this confirms why we need more robust tests. But if a length | reported is less that the actual length, and we allow the use of | 'done' tests past that point, but the 'done' is never reached ... | we're back to spam hangs :( Nuts, maybe some overrun should be | alowed: while (locallength > -100). Damn, rotten problem. | | | | ============================================================================ | | Tue, Jun 14, 2005 - 10:08:42 pm | | glennmcc> | //!!glennmcc: May 21, 2005 -- force "done" at end of message | //the next 2 lines were used during my testing | //sprintf(str,MSG_GET2, process,locallength,count,totallength,MSG_GET3 ); | //outs(str); | //if(locallength<1) done=1; | if(locallength<1 && strstr(ptr,"\n.")) done=1; | //!!glennmcc: end | | //treat special situations: | if(lastchar[1]=='.' && lastchar[0]=='\n' && Š| (ptr[0]=='\n' || ptr[0]=='\r') || | lastchar[1]=='\n' && ptr[0]=='.' && | (ptr[1]=='\n' || ptr[1]=='\r')) | { | len=0; | done=1; | } | else | { | //!!glennmcc: | // this next section had already been commented-out by Michael Polak | /* | while(len>2 && | (ptr[len-1]=='\n' || ptr[len-1]=='\r' || ptr[len-1]=='.')) | { | len--; | if(ptr[len-1]=='\n' && ptr[len]=='.' && | (ptr[len+1]=='\n' || ptr[len+1]=='\r')) | { | ptr[len]='\0'; | done=1; | break; | } | } | */ | | | if(strstr(ptr,"\n.\r") || strstr(ptr,"\n.\n")) | { | char *dot=strrchr(ptr,'.'); | if(dot) | *dot='\0'; | done=1; | } | | /* | //!!glennmcc: May 21, 2005 -- this old fix is no longer needed :) | //!!glennmcc: begin Aug 07, 2004 | //fix the 'freeze-up' problem on certain SPAM emails | //if (ptr[0]==atoi("0") && ptr[1]==atoi("10")) done=1; | { | strcpy(str,configvariable(& ARACHNEcfg,"FreezeUpString",NULL)); | if(strlen(str)>1 && strstr(ptr,str)) | { | done=1; | } | else | strcpy(str,configvariable(& ARACHNEcfg,"FreezeUpString2",NULL)); | if(strlen(str)>1 && strstr(ptr,str)) | { | done=1; | } | } | //!!glennmcc: end | */ Š| | } | | if(len>0) | lastchar[1]=ptr[len-1]; | else | lastchar[1]='\0'; | | if(len>1) | lastchar[0]=ptr[len-2]; | else | lastchar[0]='\0'; | | if(len) | write(f,ptr,len); | | thisfile+=len; | | // if(read % 512l == 0l) //kresleni tycky kazdy 1 kB | percentbar((int)(100*read/totallength)); | } | | //!!glennmcc: May 21, 2005 | //this fix for the 'freeze-up' that was happening with certain SPAM | //emails involved simply reimplementing this commented-out section | //in Michael's original line | // while (!done /*&& locallength>0*/); | //________________--->^^^^^^^^^^^^^^^^^^^^<---_____ | while (!done && locallength>0); | //in order to make 'doubly sure' those SPAMs never again cause | //that 'freeze-up'...... I also added the 'force done' section up above | //!!glennmcc: end | | //!!glennmcc: May 21, 2005 -- do not add a LF | //I don't understand why it was doing that anyways ;-) | // write(f,"\n",1); | //!!glennmcc: end | | ============================================================================ | | Wed, Jun 15, 2005 - 9:28:40 am | | ray> | Glenn, | | I dunno what's new in there :-( | | | | ============================================================================ | | Wed, Jun 15, 2005 - 12:17:51 pm | | glennmcc> | //if(locallength<1) done=1; // old force done Š| if(locallength<1 && strstr(ptr,"\n.")) done=1; // new force done | | In addition... I have also found the we to go back to simple | | while (!done); | | | | ============================================================================ | | Wed, Jun 15, 2005 - 10:07:40 pm | | ray> | Ok, will try that. | | | ============================================================================ | | Thu, Jun 16, 2005 - 9:31:08 am | | ray> | Glenn, | | It's better, but I still get aborts, maybe 1 in 5 eamils vs. about | 1 in 3 with the previous. Why not give my time-out idea a try? | make the loop wait a second or two after the pipe runs dry after | locallength is < 0 and then go on to the next mail without presuming | an error. As I see it, my aborts are happening eventhough there | is nothing really wrong. Even with the spams, the problem is that | she is aborting when she should be just coninuing the next download | after the pipe runs dry. Or is this the wrong take? | | | | ============================================================================ | | Thu, Jun 16, 2005 - 5:47:17 pm | | glennmcc> | Sounds logical to me. :)) | | ============================================================================ | | Thu, Jun 16, 2005 - 5:48:53 pm | URL | | glennmcc> | But just for the hell of it... | | Try this core.exe and see what happens. | http://www.cisnet.com/glennmcc/ara-gpl/core0615.zip | | ============================================================================ | | Thu, Jun 16, 2005 - 6:04:17 pm Š| | glennmcc> | Here's yet another one that's working pefectly here on my end. | | One test and only that one test for 'done'. and the is my newest | 'forece done' line. | (all other tests commented-out) | | //!!glennmcc: May 21, 2005 -- force "done" at end of message | //the next 2 lines were used during my testing | //sprintf(str,MSG_GET2, process,locallength,count,totallength,MSG_GET3 ); | //outs(str); | //if(locallength<1) done=1; | //!!glennmcc: Jun 15, 2005 -- trying something different | if(locallength<1 && (strstr(ptr,"\n.\n")||strstr(ptr,"\n.\r"))) done=1; | //!!glennmcc: end | | //treat special situations: | /* if(lastchar[1]=='.' && lastchar[0]=='\n' && | (ptr[0]=='\n' || ptr[0]=='\r') || | lastchar[1]=='\n' && ptr[0]=='.' && | (ptr[1]=='\n' || ptr[1]=='\r')) | { | len=0; | done=1; | } | else | */ { | //!!glennmcc: | // this next section had already been commented-out by Michael Polak | /* | while(len>2 && | (ptr[len-1]=='\n' || ptr[len-1]=='\r' || ptr[len-1]=='.')) | { | len--; | if(ptr[len-1]=='\n' && ptr[len]=='.' && | (ptr[len+1]=='\n' || ptr[len+1]=='\r')) | { | ptr[len]='\0'; | done=1; | break; | } | } | */ | | | /* if(strstr(ptr,"\n.\r") || strstr(ptr,"\n.\n")) | { | char *dot=strrchr(ptr,'.'); | if(dot) | *dot='\0'; | done=1; | } | */ | /* Š| //!!glennmcc: May 21, 2005 -- this old fix is no longer needed :) | //!!glennmcc: begin Aug 07, 2004 | //fix the 'freeze-up' problem on certain SPAM emails | { | strcpy(str,configvariable(A,"FreezeUpString",NULL)); | if(strlen(str)>1 && strstr(ptr,str)) | { | done=1; | } | else | strcpy(str,configvariable(A,"FreezeUpString2",NULL)); | if(strlen(str)>1 && strstr(ptr,str)) | { | done=1; | } | } | //!!glennmcc: end | */ | | } | | if(len>0) | lastchar[1]=ptr[len-1]; | else | lastchar[1]='\0'; | | if(len>1) | lastchar[0]=ptr[len-2]; | else | lastchar[0]='\0'; | | if(len) | write(f,ptr,len); | | thisfile+=len; | | // if(read % 512l == 0l) //kresleni tycky kazdy 1 kB | percentbar((int)(100*read/totallength)); | } | | //!!glennmcc: May 21, 2005 | //this fix for the 'freeze-up' that was happening with certain SPAM | //emails involved simply reimplementing this commented-out section | //in Michael's original line | while (!done /*&& locallength>0*/); | //________________--->^^^^^^^^^^^^^^^^^^^^<---_____ | //!!glennmcc: Jun 15, 2005 -- this fix did not work correctly after-all :( | //it caused many aborted downloads of otherwise 'normal' messages | // while (!done && locallength>0); | //in order to make 'doubly sure' those SPAMs never again cause | //that 'freeze-up'...... I also added the 'force done' section up above | //!!glennmcc: end | | //!!glennmcc: May 21, 2005 -- do not add a LF | //I don't understand why it was doing that anyways ;-) Š| // write(f,"\n",1); | //!!glennmcc: end | | | | ============================================================================ | | Thu, Jun 16, 2005 - 6:36:29 pm | | ray> | Seems to work, but does it stop spam? Nuts, I'd like to weigh in on | this problem properly but I can't since I've no spams to test. | Anyway, my hunch is that the answer will look something like this: | | if (locallength <= 0) | && ( (proper end of file dot is reached) | || (input buffer has not had new input for 2 sec) ) | done = 1; | | So I'd use locallength af first then accept either a proper end of | message or a timeout. | | | ============================================================================ | | Thu, Jun 16, 2005 - 7:26:11 pm | | glennmcc> | Will it stop the 'freeze-up spams' ??? | | I'll let you know the next time I get one. ;-) | | As to (input buffer has not had new input for 2 sec) | | Where do we set the timer to 2sec ? | | We can't do it anywhere inside of the do/while because everything inside | of it keeps getting looped-through even when no more data is coming down | the pipe. | | We can't do it outside of the do/while because the do/while will take | longer than 2sec if the current file is more than 2*bps rate. | | ============================================================================ | | Thu, Jun 16, 2005 - 7:45:56 pm | | ray> | And why do I only see that dot in pop3.log? I can see where | the dot is replaced with 0, presumably to terminate a string, but | why then is it still visible in the .log? | | | | ============================================================================ Š| | Thu, Jun 16, 2005 - 10:11:18 pm | | ray> | Ah, but that's just it, it keeps looping because it don't know better, | but if you hooked up a clock (like the ones we use for download time, | etc.) and then did something like: | | static long timeout = 0; // retain values between function calls | (linebuffer gets input); // get data | | if (something in linebuffer) // if we have data ... | timeout = 0; // all is well. | | if (nothing in linebuffer && !timeout) // no data, timer not started | timeout = (time from clock); // start the clock | | if (timeout && timeout + 3 < (time from clock) // 3 sec w. no data | { | timeout = 0; // better zero the clock | done = 1; // email is probably finished. | } | | | ============================================================================ | | Thu, Jun 16, 2005 - 10:14:44 pm | | ray> | Oh, forgot, just to avoid possible trouble, I'd wait for locallength | to zero before I'd try the above since I *assume* that the value there | is always <= the real length, and of course you could also test for | the blasted dot, to. | | | ============================================================================ | | Fri, Jun 17, 2005 - 7:24:31 pm | | glennmcc> | The dot is in the .LOG but not in the .CNM because the data gets written | into the .LOG before the dot gets converted to a 'null' | | The data is written into the .CNM _after_ the dot is converted to a 'null'. | | | | ============================================================================ | | Fri, Jun 17, 2005 - 9:09:04 pm | | ray> | Yeah, that makes good sense. BTW, I'll leave it to you to document | those Y/N defaults since you can explain them better than me. | Š| | ============================================================================ | | Wed, Jun 22, 2005 - 12:24:44 pm | | ray> | Joe, all, | | Well, I just combined the seperate .cfgs I was using for my cores | vs. Glenn's and the 'miracle' has vanished. Glenn' is now back to | being a tad faster than mine, so the problem wasn't that I had | a miracle, but that Glenn's was running slow due to some option | that's changed in the new combined .cfg. I was fooled by the fact that | all my changes made a small and proportionate change in both cores. | So the much more sencible problem is now for me to figure out why | an unpadded .cfg makes my core *slow*. And even that has vanished. | Ghosts! | | | ============================================================================ | | Wed, Jun 22, 2005 - 4:25:18 pm | | glennmcc> | Yep...... | | ghosts in the machine, and glitches in the Matrix. ;-) | | | ============================================================================ | | Wed, Jun 22, 2005 - 9:36:00 pm | | ray> | No ghosts after all. For some reason, the fast loads were from | cache, the slow without. So the question now it to figure out | how in hell padding my .cfg could result in an .ASF using cache | whereas with, it reloaded the paged from scratch. | | | ============================================================================ | | Sun, Jun 26, 2005 - 8:28:53 am | | joe> | More tales from the crypt: | | My little speed checker runs faster if I run Arachne with _less_ | free memory. 13sec. from the command line with 206K free, 9. sec | if run shelling out from BC with only 70K. Glenn's core, respective | numbers are 12 sec. 8 sec. Needless to say, this makes no sense at | all. Then there's my faster times when arachne.cfg is padded. | | Can you guy run some similar checks on your cores and see if you | can confirm these results? I'd be happier if it was some anomaly Š| on my machine. | | | ============================================================================ | | Sun, Jun 26, 2005 - 9:33:40 am | | ray> | Ooops, sorry Joe, I didn't mean to impersonate you that was me, I | meant: | | Joe, | | xxxx | | ... not to have your name on the Nick line. | | | ============================================================================ | | Sun, Jun 26, 2005 - 10:55:59 am | | ray> | Yep, something funny going on. I'm starting this #define DEBUG for | some of the stuff I've added that ain't really needed. Guess what | happens to my speed when I turn that code OFF ... speed is | reduced. This defies all reason, gotta be some mistake on my part, | but I'd like to know what it is so I can run meaningful speed tests. | Right now it looks like Glenn's core is 3% faster than mine, but | in a world where less memory == more speed, that's to be expected ;-))) | Help me with this gentlemen, it'll have me in the nut house. | Glenn, your latest experiments have you consuming less memory, see | if that degrades your performance. Next I'll check Joe's core | for speed against ours. | | | ============================================================================ | | Sun, Jun 26, 2005 - 10:15:45 pm | | gregy> | The new CORE Testers' Message Board is up, right next door to yours. | | | http://www.hi-line.net/~gfeig/ctst/ctsindex.php | | | | ============================================================================ | | Sun, Jun 26, 2005 - 10:17:40 pm | | gregy> | I'm trying to get this board coded for the new-style "space-preserving" | "indentable" form. I want to get this done before the end of 06Q2.. Š| (i.e. 30 June)... | | | ============================================================================ | | Mon, Jun 27, 2005 - 4:31:59 pm | | ray> | Recent attempt to access Yahoo to download Joe's core failed with | mine and 1.83 :( | | | ============================================================================ | | Mon, Jun 27, 2005 - 8:53:21 pm | | glennmcc> | Ah Ha !! ....... I finally got it !!!!! | | ---- popget.c --- | //!!glennmcc: Jun 27, 2005 -- This one finally works correctly :)) | if(locallength<1 && | ( | strstr(ptr,"\r\n\0\r\n.\r\n") | || strstr(ptr,"\n\0\n.\n") | ) | ) done=1; | //!!glennmcc: end | __________________________________________________________________________ | | ============================================================================ | | Tue, Jun 28, 2005 - 3:40:38 am | | Joe> | Hi all, | | I wish Michal was still with us, these strange speed/memory effects | are just the sort of thing he'd have been able to find a explanation | for! :-/ | | If what you say is correct Ray, perhaps it's a quirk of Borland's | run-time library. Perhaps to minimize memory fragmentation, each | time memory is malloc'ed or released, stuff gets moved around to | eliminate/minimize "holes". The more memory to keep tidy, the more | CPU time is used doing so. Well, that's all I can think of that | comes to mind and isn't crazy. This doesn't explain your "bloated | cfg" effect however, since that has a very distinct threshold, right? | | Also, what problem do you have trying to download my 1.83J1 core | from geocities? Note as well this is still compiled with the first | build of wattcplg.lib, using TASM 3.2, should I recompile with | my latest build of wattcplg.lib and also with Glenn's? (I still | don't know what, if any, wattcplg.lib factors affect Core's speed, | since they all behave the same for me.) Š| | | ============================================================================ | | Tue, Jun 28, 2005 - 7:29:13 am | | ray> | Joe, | | Yup, Michal would disasemble the whole thing and find the reason. | I like your theory vis. smaller==slower. More free memory could | very well require more 'management' in some way. Maybe it's something | in XSWAP. The thing with the .cfg was a distinct threshold but only | with 'screenmode != A' which is something to do with virtual memory ... | heck that could be it, those were huge pages (1500 item mail lists) | maybe if there is not enough memory she doesn't even try making a | virtual screen and runs faster! Thus more memory == slower running. | | As to the .lib thing, at the end of the day the core built with | Glenn's vs. your .lib is only different in two bytes so it can't be | that difficult to figure out what those two bytes do, I figure. | | | ============================================================================ | | Tue, Jun 28, 2005 - 7:30:35 am | | ray> | Oh, forgot ... I never tried geocities, it was yahoo that gave | trouble. Please give me the URL for your geocities site and I'll try | that | | Tx. | | | ============================================================================ | | Tue, Jun 28, 2005 - 10:57:12 pm | | gregy> | Glennmcc, et al, please try posting some code snippets into the CTST | board, now. I think it's fixed. | ... | If it works, we'll wait a couple days to shake it out over there, then | put that board up here on the 30 June, and archive all the old posts, | and go with a new-style board, here, too... | | | | ============================================================================ | | Wed, Jun 29, 2005 - 2:57:01 am | | gregy> | I think I have the DOS Testers1 Board straightened out. BUT, please do Š| not go there and bang on the code to see if you can break it. | | I put the same .PHP and SRC.TXT up in my brdTEST area at: | | http://www.hi-line.net/~gfeig/brd/test/tindex.php | | Please do all your banging and testing on that one, instead of the one | I have up in the /CTST/ area. | | On the 1 July, I will be putting that new code in place to run this | board, so you might want to check it out, now, so we know it isn't all | messed up, before we implement it here. | | | | ============================================================================ | | Wed, Jun 29, 2005 - 3:06:05 am | | Joe> | Hi Ray, et al. | | I think you mean "virtual screens", not "virtual memory" ... | | Anyway, there should be no gradation in speed with XSWAP itself, | it only has two distinct modes - single or dual XSWAP buffers, | depending on whether there is enough memory availble for one | or two buffers, of course. | | Except for the "cfg stuffing", you're describing a gradual change | in speed vs. memory, so I don't think XSWAP is responsible per se, | it must be a Borland RTL thing, I think. | | As for my (much neglected;-) web site, it's www.geocities.com/j_ds_au | If you're after the original 1.83J1 core compilation (using the TASM | 3.2 build of wattcplg.lib), you'll find it there. | | If you're after the sources, you'll find them at AD, per the | incremental list of source revisions I posted at AD (in answer | to a query from Vladimir). Else if you want the sources in one | complete archive, Vladimir collected them per the list and | uploaded this to his web site, with a message posted at AD with | the URL. Just enter '1.83J1' in the message search box at AD, | this should quickly bring up all messages related to 1.83J1, | much quicker than browsing the message archive. | | ============================================================================ | | Wed, Jun 29, 2005 - 3:11:13 am | | gregy> | Joe, you're up awfully late, tonight (morning..??)..... | | | Š| ============================================================================ | | Thu, Jun 30, 2005 - 2:36:46 am | | Joe> | Hi Gregy, | | Hmmm ... I think there's a bit of a time difference involved ... | let's see ... it's about 6:39pm local time ATM ... | | ============================================================================ | | Thu, Jun 30, 2005 - 2:39:30 am | | Joe> | Interesting ... it seems I'm 16 hours ahead of US MDT, whatever | that means (yes, I know what US and T mean;-) ... | | ============================================================================ | | Thu, Jun 30, 2005 - 3:01:25 am | | gregy> | Joe...I knew that...I was just being smart-alecky...... | | | | ============================================================================ | | Thu, Jun 30, 2005 - 3:02:57 am | | gregy> | ..and...MountainDaylightTime....I mean, I don't live in Montana, without | being on "mountain" time.... | | | ============================================================================ | | Thu, Jun 30, 2005 - 7:02:26 am | | ray> | Joe, | | Yes, 'virtual screen' it is ;-) could the padded .cfg perhaps force | her from two buffers to one? That sounds testable. | | | ============================================================================ | | Thu, Jun 30, 2005 - 8:38:20 am | URL | | gregy> | Got your new-style board up at: | Š| http://www.hi-line.net/~gfeig/brd/test/tindex.php | | You guys might want to bang on it a little before it goes into | operation | here, tonight at midnight... | | | | ============================================================================ | | Thu, Jun 30, 2005 - 7:14:23 pm | | ray> | testing | 0 | 1 | 2 | 3 | | | ============================================================================ | | Thu, Jun 30, 2005 - 8:33:05 pm | | Joe> | Ray, | | Well, padded cfg might force XSWAP to have _one_ XSWAP buffer | instead of two, but I think the decision on buffers happens | before the cfg is loaded (haven't checked). You can tell the | numbers of buffers from the colour of the "free memory" display, | green is two, black is one, red is one or zero (haven't checked). | | ============================================================================ | | Thu, Jun 30, 2005 - 8:34:56 pm | | Joe> | Hi Gregy, | | Geee, those mountains must sure be high, if they are still | in daylight at 2:30am! | | | ============================================================================ | | Thu, Jun 30, 2005 - 10:18:08 pm | | glennmcc> | Time check | | ============================================================================ | | Thu, Jun 30, 2005 - 11:52:37 pm | Š| gregy> | I'll be dropping this board down, in about 10 minutes, to archive | it and | upload the new-style board, so I'd appreciate it if y'all would | refrain | from posting for about another 1/2 hour (okay, so it takes us more | mature folks a little longer to get things done......).. | | | ============================================================================